Summary of diabetes study - DSpace at Cambridge - University of ...

The Effect of Glibenclamide on the
Pathogenesis of Melioidosis
Gavin Christian Kia Wee Koh
Emmanuel College, University of Cambridge
This dissertation is submitted for the degree of Doctor of Philosophy
31 January 2012

The effect of diabetes mellitus and glibenclamide on the
pathogenesis of melioidosis. Gavin Christian Koh Kia Wee.
This submission is my own work and contains nothing which is the
outcome of work done in collaboration except where specifically
where due acknowledgment has been made in the text. This work
was conducted at the Wellcome Trust Mahidol-Oxford Tropical
Medicine Research Unit, Mahidol University, Bangkok, Thailand
(April 2007 to November 2008), the Center for Experimental and
Molecular Medicine, Academic Medical Center, University of
Amsterdam, Amsterdam, The Netherlands (March 2009 to
December 2010), and the Wellcome Trust Sanger Institute,
Cambridge (January 2011 to December 2011) and was supervised by
Professor Sharon Jayne Peacock (Department of Medicine,
University of Cambridge and the Wellcome Trust Sanger Institute),
Professor Nicholas Philip John Day (Mahidol University &
University of Oxford), Professor Gordon Dougan (Wellcome Trust
Sanger Institute), Dr Willem Joost Wiersinga (University of
Amsterdam), and Professor Tom van der Poll (University of
Amsterdam).
This dissertation is printed single-sided on A4 paper in one-and-a-
half spaced type, as required by the Board of Graduate Studies and
the Degree Committee for Clinical Medicine and Clinical Veterinary
Medicine, University of Cambridge and does not exceed 60,000
words after excluding figures, photographs, tables, appendices and
bibliography.
The fount is 12 point Georgia, which is a transitional serif typeface
designed in 1993 by Matthew Carter and has a larger x-height to
increase clarity at smaller fount sizes. A distinctive feature of the
fount is its use of text figures.
ad maiorem dei gloriam

My father, mother & godmother
The effect of diabetes mellitus and glibenclamide on the pathogenesis of
melioidosis. Copyright © 2009–2011 Gavin Christian Kia Wee Koh
Permission is granted to copy, distribute and/or modify this document under
the terms of the Creative Commons Attribution-Noncommercial-Share Alike 2.0
UK: England & Wales Licence, provided this work is properly cited. A copy of
the licence is included in the section entitled “Creative Commons Licence”. A
copy of the licence is also available at http://creativecommons.org/licenses/by-
nc-sa/2.0/uk/legalcode.

Table of contents
List of figures .................................................................................. iii!
List of tables ..................................................................................... v!
List of abbreviations ....................................................................... vi!
Summary .......................................................................................... 1!
1.! Introduction ............................................................................... 2!
1.1! Burkholderia pseudomallei ........................................................................ 2!
1.2! Epidemiology ........................................................................................... 12!
1.3! Clinical features of melioidosis ............................................................... 20!
1.4! Diagnosis ................................................................................................. 28!
1.5! Therapy ................................................................................................... 32!
1.6! Adjunctive therapies ............................................................................... 39!
1.7! The host response to melioidosis ............................................................ 42!
1.8! Diabetes and infection ............................................................................. 59!
1.9! Diabetes and melioidosis ........................................................................ 66!
1.10! Outline of the work ................................................................................ 69!
2! Glibenclamide, not diabetes, protects from mortality in
melioidosis .................................................................................... 70!
2.1! Introduction ............................................................................................ 70!
2.2! Methods ................................................................................................... 82!
2.3! Results ..................................................................................................... 85!
2.4! Discussion ............................................................................................... 96!
3! Glibenclamide is associated with reduced inflammation in
melioidosis .................................................................................... 99!
3.1! Introduction ............................................................................................ 99!
i

3.2! Methods .................................................................................................. 107!
3.3! Results .................................................................................................... 109!
3.4! Discussion .............................................................................................. 114!
4! Glibenclamide reduces interleukin 1β secretion in a mouse model
of melioidosis ................................................................................ 116!
4.1! Introduction ........................................................................................... 116!
4.2! Methods .................................................................................................. 125!
4.3! Results .................................................................................................... 127!
4.4! Discussion .............................................................................................. 135!
5! Final discussion and conclusion ............................................... 138!
5.1! Cardiac effects of glibenclamide ............................................................ 139!
5.2! Separating the anti-inflammatory effects of glibenclamide from its
effects on insulin secretion ............................................................................. 141!
5.3! Suggestions for future research ............................................................ 144!
A! Appendix. Microarray results .................................................. 145!
References .................................................................................... 151!
Creative Commons Licence ........................................................... 221!
List of manuscripts ....................................................................... 227!
Acknowledgements ...................................................................... 228!
ii

List of figures
Figure 1. Wallace’s Line. ......................................................................................... 4!
Figure 2. Microscopic appearance of B. pseudomallei. ......................................... 5!
Figure 3. Appearance of B. pseudomallei in clinical specimens. .......................... 7!
Figure 4. Map of the global distribution of melioidosis. ..................................... 13!
Figure 5. The relationship between rainfall and melioidosis cases in Ubon
Ratchathani, Thailand. ................................................................................... 17!
Figure 6. Farming barefoot in northeast Thailand. ............................................. 18!
Figure 7. Clinical manifestations of melioidosis. ................................................ 22!
Figure 8. A 2-year-old child with unilateral parotitis at Sappasithiprasong
Hospital in Ubon Ratchathani, Thailand. ..................................................... 24!
Figure 9. Seroprevalence rises with age in Northeast Thailand. ......................... 27!
Figure 10. Histological features of human melioidosis. ..................................... 43!
Figure 11. Sappasithiprasong Hospital, Ubon Ratchathani. ................................ 71!
Figure 12. Ubon Ratchathani province. ............................................................... 72!
Figure 13. The logistic function. ........................................................................... 73!
Figure 14. The influence of diabetes on mortality is confounded by gender. ..... 77!
Figure 15. Conceptual hierarchical framework of risk factors for melioidosis
mortality. ....................................................................................................... 82!
Figure 16. Summary of patient recruitment for cohort study. ........................... 86!
Figure 17. Kaplan-Meier survival curves of 1160 patients with melioidosis. ..... 90!
Figure 18. Glibenclamide rINN or glyburide USAN. ........................................... 97!
Figure 19. General structure of sulphonylureas and sulphonamides. ................. 97!
Figure 20. Illumina BeadArray. ......................................................................... 101!
Figure 21. Differential expression of inflammation-associated genes in diabetic
patients with acute melioidosis with or without glibenclamide. ................. 112!
Figure 22. Effect of glibenclamide pre-treatment on neutrophil stimulation .. 113!
Figure 23. A. D-Glucose. B. Streptozocin. .......................................................... 122!
Figure 24. Development of hyperglycaemia in mice treated with STZ. ............ 123!
Figure 25. Glucose concentrations do not fall with glibenclamide treatment in
strepztozocin-treated mice. .......................................................................... 128!
iii

Figure 26. Bacterial loads in the lungs and BALF are unaffected by
glibenclamide therapy. ................................................................................. 129!
Figure 27. Cytokine levels in glibenclamide-treated mice (inflammasome
related). ......................................................................................................... 130!
Figure 28. Lung infiltration in glibenclamide-treated mice. ............................. 131!
Figure 29. Bacterial loads in blood, liver and spleen are lower in glibenclamide-
treated mice. ................................................................................................. 132!
Figure 30. IL18 and IFNγ levels in glibenclamide-treated mice. ...................... 133!
Figure 31. Levels of cytokine unrelated to the inflammasome in glibenclamide-
treated mice. ................................................................................................. 134!
Figure 32. Structure of glibenclamide and its metabolites. ............................... 142!
Figure 33. Structure of meglitinide. ................................................................... 143!
iv

List of tables
Table 1. Summary of Burkholderia pseudomallei virulence factors. .................. 11!
Table 2. Isolation rate of B. pseudomallei from soil compared to incidence of
melioidosis by region of Thailand. ................................................................. 16!
Table 3. Summary of seroprevalence studies conducted using IHA. .................. 27!
Table 4. Summary of antimicrobial chemotherapy of melioidosis. .................... 35!
Table 5. Susceptibility of B. pseudomallei to first-line parenteral agents. ......... 37!
Table 6. Cytokines and chemokines involved in clinical melioidosis. ................ 46!
Table 7. Alternative names for x and y in multivariable regression models. ...... 74!
Table 8. A hypothetical study population. ........................................................... 78!
Table 9. Patient characteristics, clinical features of melioidosis and outcome. . 88!
Table 10. The effect of diabetes on mortality. ..................................................... 92!
Table 11. Effect of diabetes on hypotension and respiratory failure. .................. 95!
Table 12. Comparison of gene expression study patient characteristics. ........... 111!
Table 13. Results of quantitative reverse-transcriptase polymerase chain
reaction. ........................................................................................................ 113!
v

List of
abbreviations
>, greater than.
≥, greater than or equal to.

via the classical pathway. Serum
concentrations of CRP are used as a
non-specific marker for sepsis.
C3, complement component 3.
C3H/He, a strain of inbred agouti
mouse; the strain was developed in
1920 from the DBA mouse by
L.C. Strong. There are two major
substrains: C3H/HeN (NIH) and
C3H/HeJ (Jackson Laboratories).
C4, complement component 4.
C5, complement component 5.
C57Bl/6, a strain of inbred mouse
that originated in the early 20 th
century with the breeder, Abbie
Lathrop (1868–1918) who lived in
in Granby, Massachusetts. The
strain is commonly referred to as
just ‘Black 6’ and is indeed black.
There are two important substrains:
C57Bl/6J (Jackson Laboratories)
and C57Bl/6N (NIH).
C8, complement component 8.
CAP, community-acquired
pneumonia.
CASP, caspase.
caspase, cysteine-dependent
aspartate-directed protease.
CAZ, ceftazidime.
CD, cluster of differentiation. This is
an internally recognised protocol
for immunophenotyping cells on
the basis of their surface antigens.
The CD classification includes
vii
signalling molecules and
membrane-bound receptors, but
the protocol is not designed to
capture information about function.
CD11a, cluster of differentiation 11a;
also known as integrin α L chain,
ITGAL or αL. It is an adhesion
molecule expressed by all
leukocytes and also has a role in
immune signalling. The
CD11a/CD18 complex is known as
lymphocyte function-associated
antigen 1 (LFA1).
CD11b, cluster of differentiation 11b;
also called integrin α M chain,
ITGAM or αM. The CD11b/CD18
complex is known as Macrophage
antigen 1 (Mac-1).
CD18, cluster of differentiation 18;
also called integrin β 2 chain,
ITGB2 or β2. CD18 forms complexes
with CD11a and CD11b.
CD54, cluster of differentiation 54;
see ICAM-1.
CE, common era (anno Domini).
chr., chromosome.
cfu, colony forming units.
CI, confidence interval.
CLED, cysteine lactose electrolyte
deficient agar.
co-amoxiclav, amoxicillin-
clavulanate.
co-trimoxazole, fixed combination
trimethoprim-sulphamethoxazole.

comb. nov., combinatio nova
(Latin), new combination (marking
the first instance of a binomial in
the literature).
Cpe, the mouse gene,
carboxypeptidase E.
CXCL, a family of cytokines
containing a cysteine-x-cysteine
motif, where x can be any amino
acid. The most prominent member
of this family is IL8, also known as
CXCL8.
DBA, a strain of inbred brown
mouse. This is the oldest strain of
inbred mouse in existence and
originates from Dr. C.C. Little in
1909. The substrains DBA/1 and
DBA2 separated in 1929.
DCCT, Diabetes control and
complications trial, was a
randomised-controlled trial of 1,441
volunteers with DM1 comparing
intensive insulin therapy with
standard therapy. 1 It was the
largest, most comprehensive study
of diabetes ever conducted at the
time. DM2 patients were excluded
from the study. It was funded by the
US NIDDK and subjects were
recruited from 29 centres in the US
and Canada.
DIC, disseminated intravascular
coagulation.
DM, diabetes mellitus.
viii
DM1, diabetes mellitus, type 1.
DM2, diabetes mellitus, type 2.
DMSO, dimethyl sulphoxide.
DNA, deoxyribonucleic acid.
E. coli, Escherichia coli.
e.g., exempli gratia; for example.
EDTA, ethylenediaminetetraacetic
acid.
ELISA, enzyme-linked
immunosorbent assay.
FDA, United States Food and Drug
Administration.
FDR, false discovery rate. In the
statistical analysis of high
dimensional biological data, the
false discovery rate is the
proportion of type I errors out of all
the comparisons found to be
statistically significant. The
Benjamini-Hochberg method and
its variants are commonly used to
explicitly control the FDR.
FiO2, the fractional concentration of
inspired oxygen. Normal room air
is 21% oxygen or 0·21.
g, gram or gramme.
G-CSF, granulocyte colony-
stimulating factor.
Gb, glibenclamide (rINN); glyburide
(USAN).
GLUT2, glucose transporter 2
(official name now SLC2A2).
glinide, meglitinide analogue. A
class of insulin secretagogues used

to treat diabetes, that are not
structurally related to
sulphonylureas, but which have a
similar mode of action: examples in
current clinical use include
glimepiride and repaglinide.
GM-CSF, granulocyte macrophage
colony-stimulating factor.
GO, gene ontology; an international
project to accurately annotate all
genes and gene products.
h, hour(s).
Hb, haemoglobin.
HbA1c, haemoglobin A1c
(glycosylated haemoglobin).
HDL, high density lipoprotein. HDL
deficiency occurs in Tangier
disease.
HES, hydroxyethyl starch.
HLA, human leukocyte antigen.
There are two classes. The class I
antigens are designated A, B and C.
The class II antigens are DP, DM,
DOA, DQ and DR.
HRP, horseradish peroxidase.
HUGO, Human Genome
Organisation.
ICAM-1, intercellular adhesion
molecule 1; also known as CD54.
ICAM-1 binds to the integrins
CD11a/CD18 and CD11b/CD18.
ICE, interleukin 1 converting
enzyme; now known as caspase 1.
i.e., id est; that is.
ix
IFA, indirect immunofluorescence.
IFFC, International Federation of
Clinical Chemistry.
IFNγ, interferon gamma.
IgG, immunoglobulin G.
IgM, immunoglobulin M.
IHA, indirect haemagglutination
test.
IL, interleukin. The interleukins are
small messenger proteins that play
a role in leukocyte signalling.
IPM, imipenem; a carbapenem
antibiotic active against
B. pseudomallei.
IUPAC, International Union of Pure
and Applied Chemistry. The
international authority for
standards in the naming of the
chemical elements and their
compounds.
KATP-channels, ATP-sensitive
potassium channels.
KC, keratinocyte-derived cytokine;
now called CXCL1.
KEGG, Kyoto Encyclopedia of Genes
and Genomes; an online database
of protein and pathway information
hosted by the Kyoto University.
KIR2DL1, killer cell
immunoglobulin-related re;ceptor,
2 domains, long cytoplasmic tail, 1.
kPa, kilopascals; see Pa.
L, litre.

L. pneumophila, Legionella
pneumophila.
LAMP, loop-mediated isothermal
amplification.
LB, lysogeny broth; also called Luria-
Bertani medium, or Luria broth.
LD50, median lethal dose. The dose
sufficient to kill half the population
tested.
LFA-1, lymphocyte function-
associated antigen 1; see CD11a.
LIX, lipopolysaccharide-induced
CXC chemokine. now known as
CXCL5.
LPS, lipopolysaccharide; also known
as a endotoxin. A component of the
outer membrane of all Gram-
negative bacteria.
LY96, lymphocyte antigen 96; also
known as MD2. LY96 associated
with Toll-like receptor 4 and is
involved in the recognition of
lipopolysaccharide.
M, molar; 1 litre of a 1 molar solution
contains 1 mole of solute.
M1, 4-trans-hydrocyclohexyl
glibenclamide; a metabolite of
glibenclamide
M2, now called M2b.
M2a, 4-cis-hydrocyclohexyl
glibenclamide; a metabolite of
glibenclamide.
x
M2b, 3-cis-hydrocyclohexyl
glibenclamide; a metabolite of
glibenclamide.
M3, 3-trans-hydrocyclohexyl
glibenclamide; a metabolite of
glibenclamide.
M4, 2-trans-hydrocyclohexyl
glibenclamide; a metabolite of
glibenclamide.
M5, ethyl-hydroxy glibenclamide; a
metabolite of glibenclamide.
Mac-1, macrophage antigen 1; see
CD11b.
MAQC, MicroArray Quality Control
project; an initiative of the FDA to
monitor the reproducibility of
microarrays, with the view that they
will eventually be used for
diagnostic applications.
max., maximum.
MD2, see LY96.
MDP, muramyl dipeptide, a
peptidoglycan constituent of both
Gram-positive and Gram-negative
bacteria.
MEM, meropenem; a carbapenem
antibiotic active against
B. pseudomallei.
MERTH, Melioidosis Eradication
Therapy/Thailand; a multicentre
randomised controlled trial in
Northeast Thailand comparing co-
trimoxazole therapy against co-
tromoxazole and doxycycline

combination therapy due to report
at the end of 2011.
MIAME, Minimum Information
About a Microarray Experiment; a
published checklist for the
minimum information that should
be reported for all published
microarray experiments.
MIC, minimum inhibitory
concentration; the minimum
concentration of antibiotic required
to inhibit bacterial growth.
MIP-2, macrophage inflammatory
protein 2. Now known to consist of
two distinct cytokines, CXCL2 and
CXCL3.
MLST, multi-locus sequence type.
mM, millimolar. See M.
mmHg, millimetres of mercury.
1 mmHg = 133.3 Pa; 1 atmosphere
= 760 mmHg.
MODY, maturity-onset diabetes of
the young.
mol, mole. And SI unit defined as
the number of elemental entities in
12 g of carbon-12. This number is
called Avogadro’s constant, and has
a measured value of 6·022 141 79 ×
10 23 per mol.
mRNA, messenger RNA.
MSD, Merck Sharp & Dohme, a
pharmaceutical company based in
Whitehouse Station, New Jersey,
xi
USA. It manufactures and markets
imipenem/cilastatin (Primaxin®).
NCBI, National Center for
Biotechnology Information; a part
of the National Library of Medicine
in the United States.
NEQAS, National External Quality
Assessment Service. A UK body
charged with independently
auditing standards in laboratories.
NET, neutrophil extracellular trap.
NF-κB, nuclear factor κB.
NK cell, natural killer cell; a form of
cytotoxic lymphocyte.
NLR, Nod-like receptor. A family of
cytosolic pattern recognition
receptors. Unlike the TLRs, they are
not membrane-bound. Since 2008,
their new name has been
‘nucleotide-binding oligomerization
domain, leucine rich repeat’, a
name chosen to retain the same
abbreviation, while adding
descriptive information about their
molecular properties.
NIDDK, United States National
Institute of Diabetes and Digestive
and Kidney Diseases.
NIH, National Institutes of Health; a
loose federation of government-
funded research institutes in
Bethesda, Maryland.
Nod, nucleotide-binding and
oligomerization domain.

NOD, non-obese diabetic mouse.
OR, odds ratio.
p-value, probability value calculated
from a statistical test.
Pa, pascals (SI unit). 1 mmHg =
133.3 Pa; 1 atmosphere
= 101.3 kPa.
PAI-1, plasminogen activator
inhibitor-1.
PAMP, pathogen-associated
molecular patterns; these are highly
conserved antigens present on a
wide range of pathogens (viruses,
bacteria and parasites).
PaO2, partial pressure of oxygen in
arterial blood, usually expressed in
mmHg, but also measured in kPa.
PC, protein C.
PBS, phosphate-buffered saline.
PCR, polymerase chain reaction.
PEG, polyethylene glycol; also called
poly(ethylene oxide) or
polyoxyethylene.
P:F or P/F ratio, the ratio of PaO2
(in mmHg) to FiO2 (as a fraction).
The ratio is used to quantify
pulmonary function. In healthy
adults, values are usually
>380 mmHg; the cut-off for ALI is
≤300 mmHg and the cut-off for
ARDS is ≤200 mmHg.
PRR, pattern recognition receptors;
the two main families of PRRS are
the TLRs and the NLRs.
xii
PYCARD, old name for ASC.
qds, quater die sumendus; to be
taken four times a day, or every six
hours.
qPCR, quantitative reverse-
transcriptase polymerase chain
reaction; a method for quantifying
DNA.
qqh, quaque quarta hora; to be
taken every four hours.
R. argentiventor, Rattus
argentiventor.
R. exulans, Rattus exulans.
R. rattus, Rattus rattus.
R. tionamicus jalorensis, Rattus
tionamicus jalorensis.
RAW264.7, a cell line originating
from virus-transformed mouse
monocytes. The cell line sheds
murine leukaemia virus and should
therefore be handled under
biosafety II conditions.
rINN, Recommended International
Nonproprietary Name. Drug name
assigned by the WHO.
RNA, ribonucleic acid.
RNS, reactive nitrogen species; these
are ·NO, ·NO2, N2O3, ONOO – ,
ONOOH and ONOOCO2 – .
ROS, reactive oxygen species; these
are singlet oxygen, ·OH, H2O2,
O=C(O·)O – , O2 – and O2 2– .
RPMI 1640, Roswell Park Memorial
Institute medium 1640. A

commonly used cell culture
medium.
rRNA, ribosomal RNA.
S. enterica, Salmonella enterica.
S. Typhi, Salmonella enterica var
Typhi, the causative agent of
typhoid fever in humans.
S. Typhimurium, Salmonella
enterica var Typhimurium. A
model organism that causes a
typhoid fever-like illness in mice.
S. aureus, Staphylococus aureus.
sFLT-1, soluble fms-like tyrosine
kinase-1. A marker for endothelial
activation.
SLC2A2, solute carrier family 2,
member 2 (also called GLUT2).
SNK, Student-Newman-Keuls. This
is a multiple comparison method
and is discussed in the chapter on
statistical analyses. Also known as
the Newman-Keuls test.
SOP, standard operating procedure.
SPP1, Secreted phosphoprotein 1.
statin, common name for HMG-CoA
reductase inhibitors; these are
drugs used for the treatment of
hypercholesterolaemia.
STZ, streptozocin or streptozotocin.
T-cell, T-lymphocyte; a lymphocyte
that originates from the thymus.
tds, ter die sumendus; to be taken
thrice daily, or every eight hours.
xiii
THP-1, a human monocytic
leukaemia cell line.
TLR, Toll-like receptor; a class of 13
single membrane-spanning
leukocyte receptors that recognise
PAMPs. Recognition by TLRs
results in activation of the innate
immune response.
TMB, 3,3%,5,5%-
tetramethylbenzadine. A substrate
for horseradish peroxidase. TMB is
a colourless compound, but is
oxidised by horseradish peroxidase
to form a blue product that may be
detected spectrophotometrically.
TNF, tumour necrosis factor. Usually
refers to TNFα, but may refer also
to the TNF family of cytokines.
TNFα, tumour necrosis factor, alpha;
a pro-inflammatory cytokine.
TSA, trypticase soy agar.
TTS, type-three secretion system.
UGDP, University Group Diabetes
Program, was a prospective
randomised-controlled trial of the
treatment of DM2 that ran from
1960 to 1975, and published its final
report in 1982. It produced two
controversial observations on the
safety and efficacy of two oral anti-
diabetic treatments: first, that
tolbutamide is associated with an
excess cardiovascular mortality, 2
and second, that phenformin is

associated with an excess mortality
from lactic acidosis. 3
UKPDS, United Kingdom
Prospective Diabetes Study, was a
randomised-controlled trial funded
by the British Diabetic Association
at a cost of £2 million. Starting in
1977, it recruited over 5,000 white
patients with recently diagnosed
type 2 diabetes and lasted for
20 years. 4
xiv
uPAR, urokinase-type plasminogen
activator receptor.
US or USA, United States of
America.
USAN, United States Adopted
Name, assigned by the USAN
Council (co-sponsored by the AMA,
USP and APhA).
USP, United States Pharmacopeial
Convention.
viz., videlicet; namely.
WHO, World Health Organization.

Summary
Melioidosis is an important cause of community-acquired sepsis, endemic
to Southeast Asia and Northern Australia. Melioidosis is caused by the soil
saprophyte, Burkholderia pseudomallei, a motile Gram-negative bacillus, and is
associated with a mortality rate that approaches 50% in Northeast Thailand.
The most important risk factor for melioidosis is diabetes mellitus, and two-
thirds of all adult patients with melioidosis have diabetes as a risk factor. It has
been noted previously, however, that patients with diabetes have lower
mortality than patients without diabetes. In this dissertation, we look at a
cohort of 1160 consecutive adult melioidosis patients presenting to
Sappasithiprasong Hospital in Ubon Ratchathani, Thailand, 410 (35%) of whom
were diagnosed with diabetes prior to admission. We confirmed previous
findings that diabetes protected from mortality in melioidosis, but also found
that this protective effect was confined to a smaller subset of patients (208
patients) who were treated with glibenclamide prior to admission. Patients with
hyperglycaemia (but no diagnosis of diabetes prior to admission) had the same
mortality rate as patients without diabetes. In vitro experiments found no
inhibitory effect of glibenclamide on bacterial growth, and we therefore looked
for evidence of an effect of glibenclamide on the host. We conducted a gene
expression study of circulating blood leukocytes in melioidosis patients and
compared them to uninfected controls. In this study, we found that
glibenclamide was associated with an anti-inflammatory effect on the host
response to melioidosis. To further elucidate a mechanism for the action of
glibenclamide, we studied the effect of glibenclamide therapy in a mouse model
of melioidosis and found that the effect of glibenclamide was specific to
interleukin-1β secretion. This reduction in interleukin-1β secretion was
associated with reduced cellular influx into the lungs as well as lower bacterial
loads in blood, liver and spleen.
1

1. Introduction
Melioidosis is infection caused by the Gram-negative bacillus,
Burkholderia pseudomallei, 5 and was first reported in 1912 in a post mortem
series of 38 patients seen by Whitmore and Krishnaswami at the Rangoon
General Hospital, Burma, the majority of whom were morphine addicts. 6–8
Whitmore recognised that the causative agent of melioidosis was a novel
organism, which he named Bacillus pseudomallei. 8 The clinical disease only
acquired its name in 1921, being coined from the Greek melis (distemper of
asses) and eidos (resemblance), and refers to its resemblance to glanders, an
infection of equids caused by the related organism, B. mallei. 9
The history of melioidosis research is punctuated by waves of activity that
coincide with the military interests of the Western powers. British interest in the
disease coincided with British imperial interests (specifically Burma for
petroleum, Malaya for petroleum and rubber), 10,11 and waned as the empire
waned. France reported approximately 100 cases among the 400,000 troops of
the French Expeditionary Forces in Indochina between 1948 and 1954. 12
America’s interest in (and therefore, research funding for) melioidosis increased
and declined proportionate to its involvement in the Vietnam war. 13–18
Following the events of September 11, 2001, funding for melioidosis research in
the last ten years has again increased with the realisation that B. pseudomallei
is easily obtained (it is an environmental organism) and when aerosolised,
might see utility as a bioterror agent with a high mortality rate.
1.1 Burkholderia pseudomallei
1.1.1 Taxonomy
B. pseudomallei was for a long time placed in the genus Pseudomonas on
the basis of its Gram-stain appearance and biochemistry 19 (it is a non-lactose
fermenting oxidase-positive Gram-negative bacillus 20 ), but the organism was
moved to a new genus, Burkholderia, in 1992 on the basis of 16S ribosomal
RNA (rRNA) sequence. 21
The causative organism of melioidosis and its characteristics were
described very shortly after the disease was first described. 8 Initially named
2

Bacillus pseudomallei by its discoverer, 6 it was later named Bacterium
whitmori (Stanton & Fletcher 1921) for its discoverer, by which appellation it is
still known to French researchers (bacille de Whitmore or Whitmore’s
bacillus). 22
In common with many other non-fermenting Gram-negative bacteria, the
organism has struggled to find a taxonomic home. The 5 th edition of Bergey’s
Manual placed it in the genus Malleomyces 23 due to the many characteristics it
shares with the causative agent of glanders (B. mallei was type species for the
now-defunct genus Malleomyces). 24,25 It was, for a time, called Pfeifferella
whitmori. 26 It was suggested also that the bacterium be housed in the genus
Loefflerella. 27 The organism was transferred to the genus Pseudomonas in the
6 th edition of Bergey’s Manual, 19 a decision supported by the best
morphological, phenotypic and biochemical evidence available at the time. 28
The organism was finally transferred to the genus Burkholderia in 1992 (as
Burkholderia pseudomallei comb. nov.) on evidence from its 16S ribosomal
subunit sequence: 21 B. pseudomallei and P. aeruginosa are now no longer even
in the same class, let alone the same genus, family or order. B. mallei was also
moved to the same genus using the same criteria, 21 and has been shown to have
evolved from B. pseudomallei by specialisation to an equine host and a process
of genome reduction. 29 Indeed, B. mallei may be considered a subspecies of
B. pseudomallei by multilocus sequence typing. 30 Although B. mallei can
survive in tap water for up to four weeks (and horses can become infected from
contaminated drinking water), 20 B. mallei has lost the ability to persist
indefinitely in the environment.
In environmental isolates, the main differential for B. pseudomallei is
B. thailandensis. One puzzle of the early 1990’s was that environmental surveys
of B. pseudomallei seemed to show a much larger endemic area than was
clinically evident. 31,32 Many of those environmental isolates were later shown to
be due to a separate species, B. thailandensis, which is distinguished from
B. pseudomallei on the basis of its ability to assimilate arabinose and its
avirulence in mice and hamsters. 32–37 In 1998, B. thailandensis was given status
as a separate species on the basis of evidence from its 16S ribosomal subunit
sequence. 38
3

Phylogenetic studies combined with geographical data support an
Australian origin for B. pseudomallei with a single introduction event into
Southeast Asia hypothesised to have occurred during a recent glacial period.
B. pseudomallei isolates may be resolved into two distinct populations divided
by Wallace’s Line 39 (a deep-water channel that forms a physical boundary
separating Australian flora and fauna from those of Asia). Wallace’s Line has
separated Australia and Asia for a period in excess of 50 million years, even
during glacial advances (when sea levels dropped by up to 120 metres from
current levels). It runs along the Lombok Straits, separating Bali (to the west)
and Lombok (to the east). Further north, Wallace’s Line runs between Borneo
(west) and Sulawesi (east).
Figure 1. Wallace’s Line.
Wallace’s Line marks the end of the Sunda Shelf, an area of shallow ocean linking the
islands of Sumatra, Java, Bali and Borneo to the continent of Asia. Lydekker’s Line marks
the limit of the Sahul Shelf, which links Australia and New Guinea. Weber’s line marks
the transitional area between the two. Image ©2007 Maximilian Dörrbecker.
4

1.1.2 Laboratory identification of B. pseudomallei
The clinical features of melioidosis are non-specific, and the diagnosis of
melioidosis therefore depends on finding B. pseudomallei in a clinical
specimen. 40,41 It is recommended that B. pseudomallei be handled in biosafety
level 3 facilities 42 and this recommendation is motivated by the fact that it may
be acquired by inhalation and mortality following infection is high. It has been
argued that level 2 facilities offer adequate protection as long as samples are
centrifuged in sealed cups and opened only in a biosafety cabinet. 43,44
Figure 2. Microscopic appearance of B. pseudomallei.
Drawing of B. pseudomallei in a pus from an abscess in a guinea pig. The figure is taken
from Stanton & Fletcher’s 1932 monograph and emphasises the safety-pin appearance of
the bacterium.
B. pseudomallei is a motile, non-lactose fermenting, oxidase-positive,
Gram-negative bacillus. 19 Although initially believed to be a strict aerobe, 20
B. pseudomallei will grow anaerobically if nitrate is present as a terminal
electron acceptor. 45,46 The optimum temperature for B. pseudomallei is 24–
32°C, but it will grow at temperatures up to 42°C. 13,47 Clinical samples are
usually incubated at 37°C in 5% carbon dioxide.
5

B. pseudomallei is not fastidious and grows readily on a wide variety of
media, including blood agar, tryptic soy broth, LB, CLED and MacConkey. It
will even survive for extended periods in nutrient deficient media. Stanton &
Fletcher (1932) reported that B. pseudomallei would survive for 44 days in tap
water, Miller et al. (1948) reported 8 weeks. 20 Wuthiekanun has found that
B. pseudomallei can survive in distilled water for 16 years and possibly
longer. 48–50
In culture, the B. pseudomallei has a characteristic bipolar (‘safety pin’)
appearance when stained for light microscopy (Figure 2), which was first noted
by Whitmore, 8 and has been ascribed to the presence of large lipid vacuoles in
the centre of the cell that do not stain. 20 Although characteristic, 17 this safety pin
appearance is not sufficiently specific as to be diagnostic 51 and does not occur
frequently in clinical specimens. 52 B. pseudomallei does not form spores. 8
Cultured bacteria are ~0·8 × 1·5 µm, but bacteria found in clinical samples may
also form long tangled filaments (Figure 3). 41,52
Recognition of B. pseudomallei is seldom a problem for the microbiologist
when the specimen has been taken from an otherwise sterile site such as blood,
or pus from a deep collection. The main technical problem is that the doubling
time for B. pseudomallei is roughly two hours, which is longer than many other
bacteria. In specimens taken from non-sterile sites (e.g., catheter urine, sputum
and throat swabs) colonies of B. pseudomallei are easily overgrown. 53 For this
reason, selective media should be used to isolate B. pseudomallei from these
specimens. 53
Ashdown agar is made from a base of trypticase soy agar (TSA): it contains
crystal violet to inhibit the growth of Gram-positive bacteria and gentamicin to
inhibit the growth of other Gram-negatives. 53 B. pseudomallei colonies absorb
neutral red, which acts as an indicator and helps to distinguish B. pseudomallei
from other bacteria. The addition of glycerol causes the colonies to wrinkle
(colonies are smooth on TSA) and the appearance of B. pseudomallei colonies
on Ashdown agar has been likened to the appearance of cornflowers (Figure 3).
Colonial morphology is extremely variable, with seven morphotypes described,
and it is common for pure cultures of a single strain to exhibit multiple
morphologies. 54
6

A
C
Figure 3. Appearance of B. pseudomallei in clinical specimens.
Note.—Both microscope specimens are from the urine of a 71-year-old woman with
B. pseudomallei septic arthritis of the right knee. She had no urinary symptoms and she
had previously had melioidosis two years previous to this presentation. Long filamentous
bacteria were seen on Gram stain of urine (A, magnification ×1000) and
immunofluorescence staining showed these to be B. pseudomallei (B, magnification
×1000). The cells visible in the background are a combination of leukocytes and
epithelial cells. Culture of this sample confirmed B. pseudomallei (C, Ashdown agar). The
lavender colonies take on their characteristic wrinkled appearance after four-days’
incubation and are said to resemble cornflowers (D). Images A, B, C © 2008 Gavin Koh.
Image D © 2007 J Stegman.
Ashdown agar is cheap and easily prepared using readily available
ingredients. However, many laboratories in the West have moved away from
preparing their own media to using commercially prepared media, and
Ashdown agar is not commercially available. For these laboratories,
7
B
D

commercially prepared Burkholderia cepacia agar or Pseudomonas cepacia
agar are suitable alternatives. 55,56
Diagnosis may be achieved with biochemical methods. Using the API 20E
system (bioMérieux, Marcy l’Etoile, France), most B. pseudomallei isolates have
one of the following seven digit profiles: 2006727, 2206706, 2206707 and
2206727. 36 Profiles on API 20NE are 1156576, 1156577, 1556576 and 1556577. 57
Of the automated systems available commercially, VITEK 1 (bioMérieux)
correctly identifies 99% of isolates. 58 Performance on VITEK 2 (bioMérieux) has
been uniformly poor by contrast. In an initial assessment performed in 2002,
VITEK 2 correctly identified only 19% of isolates. 58 In 2006, using a new
colorimetric GN card, this had improved to only 63–81%. 59 Common
misidentifications include Burkholderia cepacia, Pseudomonas aeruginosa,
Chromobacterium violaceum and other non-lactose fermenting Gram-negative
bacilli.
Variation in the API profile occurs primarily in citrate utilisation, and in
sucrose and amygdalin assimilation. The antibiotic susceptibility pattern for
B. pseudomallei is also helpful in allowing it to be differentiated from
P. aeruginosa, because B. pseudomallei is invariably resistant to gentamicin
and colistin, but is usually susceptible to co-amoxiclav, 36,40,60 a resistance
pattern that does not occur in P. aeruginosa.
The ability of patient sera to agglutinate suspensions of B. pseudomallei
was first reported in 1932, first by the English physicians, Stanton and Fletcher,
and by the Dutch physicians, de Moor and van der Walle. Both these groups
evaluated this technique as a means of diagnosing melioidosis serologically. It
was Miller in 1947, 20 who first had the insight that the ability of serum from
infected rabbits to agglutinate unknown bacteria obtained from cultures might
be used as a means to identify B. pseudomallei. Unfortunately, Miller’s slide
agglutination test never entered routine clinical use.
The sensitivity of slide agglutination test may be increased by conjugating
the antibody to latex beads, so that even bacteria from a single colony may form
a visible precipitate. A non-commercial latex agglutination test based on a
monoclonal antibody to the exopolysaccharide of B. pseudomallei has been
clinically evaluated and has a sensitivity of 99·5%. The test has been available in
Thailand since 2002 (Chulaborn Research Institute, Bangkok, Thailand). 57,61 As
8

with previous tests based on serological methods (including Miller’s test 20 ), the
test cannot distinguish B. mallei from B. pseudomallei.
Immunofluorescence microscopy is useful in samples with a high bacterial
load 62 (≥10 4 cfu/ml; e.g., sputum, urine or pus from a normally sterile site) and
is capable of giving same-day results with >99% specificity. 63,64 Bacterial loads
in blood are too low for immunofluorescence microscopy, 65,62 which means that
immunofluorescence is not useful in patients who are bacteraemic but with no
identifiable focus. The sensitivity is reported to be 66–73% in an idealised
research setting, 63,64 but is probably lower in routine clinical practice.
Molecular techniques have been attempted, but are not routinely available
in clinical practice. Targets for polymerase chain reaction (PCR) have included
the 16S ribosomal subunit, 66,67 the gltB and narK genes 68 (which are also part of
the MLST scheme); orf11, orf13 and BpSCU2 (components of the type 3
secretion system), 69 fliC (which codes for flagellin), 67 and repetitive elements of
non-protein-coding DNA. 70 Few of these have been validated in the clinical
setting, and the need for equipment capable of PCR make the majority of these
methods impractical in the field. On top of this, mere amplification of genes
may not be sufficient and additional tests such as gene sequencing may be
required. 71
Loop-mediated isothermal amplification (LAMP) appears to have a
number of characteristics that make it preferable to PCR for nucleic acid
amplification in resource-poor settings. The method employs a DNA polymerase
and a set of four specially designed primers that recognise a total of six distinct
sequences on the target DNA. The reaction is simple to set up and requires
much less training than PCR, and the initial capital outlay is minimal (the only
equipment required is a water bath or heat block capable of maintaining a
constant temperature of 60–65°C). A LAMP assay was developed targeting the
BPSS 1406 hypothetical protein, situated within the TTS1 gene cluster of
B. pseudomallei. 72 The lower limit of detection was 38 genomic copies per
reaction, which is still too high to reliably detect B. pseudomallei in blood. 62,65
Performance was better in other sample types including sputum, urine and pus
(sensitivity 45·8–75%, specificity ≥97·9%). No direct comparison was made
against immunofluorescence, but these results seem comparable to that
9

technique, 63,64 which means LAMP is unlikely to supplant immunofluorescence
as a method for rapid antigen detection.
Techniques for rapid diagnosis are urgently needed, because culture-based
methods entail a minimum turnaround time of 48 hours or more.
Immunofluorescence microscopy currently fulfils that function, but is
insufficiently sensitive.
B. thailandensis has a biochemical phenotype very similar to
B. pseudomallei, but B. thailandensis almost never causes disease in humans.
There are only two reported cases of human B. thailandensis infection in the
literature 73,74 and the need for a clinical laboratory to distinguish between the
two species therefore almost never arises. If necessary, B. thailandensis can be
differentiated from B. pseudomallei by its ability to assimilate arabinose, which
B. pseudomallei cannot do. 34,38 A definitive distinction between the two species
may be made using MLST. 30
B. mallei (the causative agent of glanders) is morphologically and
biochemically indistinguishable from B. pseudomallei, except that the
bacterium is immotile. Like B. pseudomallei, automated laboratory methods
may misidentify it as other non-lactose fermenting Gram-negative bacilli. 75 For
a clinical laboratory on the ‘front lines’, the first indication that the isolate is
B. mallei and not B. pseudomallei is the susceptibility profile: ~80% of B. mallei
isolates are susceptible to gentamicin and may also be susceptible to some
macrolides. 76,77 Some isolates may also be susceptible to ampicillin, which
B. pseudomallei never is. 76 A definitive diagnosis may require 16S RNA
sequencing or MLST, but this issue is unlikely to arise, except in a biothreat
setting, since glanders has been largely eradicated and the last naturally
occurring infection reported in the English-language literature was in 1949. 75
Outbreaks continue to occur in susceptible equids, the most recent of which was
an outbreak in dromedaries in Bahrain. 78
1.1.3 A brief survey of virulence factors
B. pseudomallei is a facultative intracellular pathogen 79–81 and much of the
literature on host-pathogen interactions centres on, or needs to be interpreted
in light of, this fact.
10

Virulence factor Function References
BimA Actin polymerization,
intracytoplasmic movement and cellto-cell
spread
BPSS1823 Intracellular survival, swarming
motility, resistance to low pH,
protease production
Capsular polysaccharide Resistance to serum lysis; complement
resistance
Flagellin Motility; cell invasion
Lipopolysaccharide Complement resistance;
endotoxin
Quorum sensing Downstream virulence factors as yet
unidentified
11
Kespichayawattana 2000, 82
Breitbach 2003, 83 Stevens 2005, 84
French 2011 81
Norville 2011 85
Reckseidler 2001, 86 Atkins 2002, 87
Reckseidler-Zenteno 2005, 88
Warawa 2009, 89
Wikraiphat 2009, 90
Reckseidler-Zenteno 2010 91
Brett 1994, 92 DeShazer 1997, 93
Chua 2003, 94 Inglis 2003, 95
Chuaygud 2008 96
Matsuura 1996, 97 Luc[1
DeShazer 1998, 98
Wikraiphat 2009, 90 Novem 2009 99
Ulrich 2004, 100 Juhas 2004, 101
Valade 2004, 102 Song 2005, 103
Diggle 2006, 104
Lumjiaktase 2006 105
Superoxide dismutase Resistance to killing by superoxide Vanaporn 2011 106
Type IV pilin Adhesion to respiratory epithelium
Type III secretion
systems
Type VI secretion
systems
Injecting effector proteins into the host
cell; escape from endocytic vacuoles
Injecting effector proteins into the host
cell
Table 1. Summary of B. pseudomallei virulence factors.
Essex-Lopresti 2005 107
Stevens 2002, 108 Stevens 2004, 109
Warawa 2005, 110 Pilatz 2006, 111
Burtnick 2008, 112 Sun 2010, 113
French 2011 81
Schell 2007, 114 Shalom 2007, 115
Burtnick 2011, 116 French 2011 81
In 1990, Pruksachartvuthi et al. showed by electron microscopy that
B. pseudomallei was able to persist within both neutrophils and monocytes
collected from healthy human volunteers. 79 That the intracellular bacteria were
viable and capable of replication was confirmed by culture. 79 B. pseudomallei is
also capable of parasitizing a number of phagocytic and non-phagocytic cell
lines, including A549 (human adenocarcinomic basal epithelial lung cells), 80,85
CHO (Cricetulus griseus ovary cells), HeLa (human cervical epithelial cells), 80
J774.1A1 (Mus musculus BALB/c peritoneal macrophage), 117 Vero

(Cercopithecus aethiops kidney epithelial cells), 80 and U937 cells (human
monocyte line). 80 It will also invade rat alveolar macrophages 80 and
Acanthamoeba species. 118
After entry into the cell, B. pseudomallei is able to evade intracellular
killing, 85,119 lyse the endosome membrane and enter the cytoplasm. 81 In the
cytoplasm, it is able to mobilise by polymerising actin at one pole of the
bacterial cell to form membrane protrusions, which allow it to invade adjacent
cells. 82–84 Virulence factors of B. pseudomallei are summarised in Table 1.
1.2 Epidemiology
1.2.1 Geographical distribution of melioidosis
Descriptions of the epidemiology of melioidosis are hampered by the lack
of clinical awareness and laboratory facilities in areas where it is likely to be
endemic. 120,121 Without the ready availability of microbiological diagnosis,
melioidosis may go completely unrecognised even in areas where it is highly
endemic. 120 The geographical distribution of B. pseudomallei should therefore
be regarded as poorly defined.
The first evidence for melioidosis in Thailand was from a 1930 case report
of a Russian man, normally resident in Bangkok, who presented in a delirious
state to police in Phnom Penh, Kampuchea (now Cambodia). Diagnosis was
made post mortem. 122 The first case of melioidosis reported in Thailand itself
was in 1947, in a Dutch soldier who had been held a prisoner-of-war by the
Japanese. 123 The first reported case in a native Thai had to wait until 1955. 124 It
was not until the expansion of microbiological services that occurred in the
1970’s and 80’s that it became clear melioidosis was highly endemic in Thailand
and probably always had been. 125,126
The opposite has occurred in Burma. In 1917, Krishnaswami reported over
a hundred cases of melioidosis in Rangoon just five years after first describing
the disease. 127 Yet, no case of indigenous melioidosis was reported in Burma
between 1945 40 and 2000 128 despite the fact that Burma continued to export
cases of melioidosis throughout that period 129,130 and despite a high
seroprevalence in the native population. 131
The global distribution of melioidosis has been described in detail by
Currie, Dance and Cheng, 132,133 and their map is reproduced in Figure 4.
12

.
Figure 4. Map of the global distribution of melioidosis.
Note.—Pink asterisks show the location of three outbreaks that occurred in temperate
regions: one in France, and two in Australia. Figure taken from Cheng and Currie Clin
Microbiol Rev 2005;18:383–416.
Melioidosis is a disease of the tropics and appears to be confined to a band
roughly 20° north of the equator and 20° south. The endemic area is focused
around southeast Asia and northern Australia; its northernmost extent is
southern China, Hong Kong and Taiwan; its southernmost extent is northern
Australia. Melioidosis is almost certainly endemic to India 134–136 and probably
always has been. There are sporadic cases from Africa and Central
America, 132,133,137,138 but the lack of microbiological facilities in much of these
areas means the true extent of melioidosis is unknown.
1.2.2 Melioidosis as a zoonosis
A zoonosis is a disease where infection occurs by transmission from a
vertebrate animal to human (World Health Organization, Geneva, Switzerland).
In 1925, Stanton and Fletcher described melioidosis as a zoonosis based on
13

observations that rodents succumb to the disease. 10 However, in that same
paper, Stanton and Fletcher also noted that melioidosis is fatal to rodents,
which makes it very unlikely that rodents could be an efficient reservoir or
vector. In their 1932 monograph, they revised this assertion, stating, ‘The way in
which man contracts melioidosis is a problem as yet unsolved,’ but continued to
assert that they ‘believe that infection usually occurs…through the consumption
of food which has been contaminated by the excreta of infected rodents.’
In 1948, Harries et al. noted that they were unable to find melioidosis in
any of >500 rats examined in Rangoon, 129 and in 1969 Strauss et al. performed
a similar survey of 422 wild rats (Rattus tionamicus jalorensis,
R. argentiventor, R. exulans and R. rattus) in a survey of Carey Island,
Selangor, Malaysia, 140 and expressed doubt that rats are a reservoir for the
disease. In the 1955, Chambon et al. showed that B. pseudomallei is primarily
environmental, 141 and transmission from animals to humans has not been
shown to be an important cause of infection. 132
The disease is dealt with by the Bacterial Zoonoses Branch of the US
Centers for Disease Control and melioidosis continues to be listed with the
zoonoses by textbooks. 142 Some books published as recently as 2002 continue to
assert erroneously that animals are a reservoir for disease. 143
Melioidosis infects a large numbers of animals and birds. 144 Sheep and
goats are particularly susceptible, resulting in the absolute requirement for
pasteurisation of tropical commercial goat’s milk. 144 Gorillas are also extremely
susceptible, so much so that extraordinary measures are needed to isolate them
from the environment in a melioidosis-endemic area. 145 Four male animals
imported to the Singapore Zoological Gardens died within months of arriving
from Rotterdam, and the one surviving animal was returned to Rotterdam for
treatment. On outbreak at the Paris zoo in the 1970’s («L’affaire du jardin des
plantes») was attributed to an imported panda. 146 Horses were initially reported
to be immune, as demonstrated by attempts at experimental inoculation, 10 but
cases of naturally occurring infection have subsequently been reported. 146–149
There appears to be no basis for the oft-repeated 150 claim that water buffalos are
immune.
Rats, mice and hamsters are all susceptible and are important small
animal models of melioidosis. 151–153 Of the three, Syrian golden hamsters are the
14

most susceptible. 154,155 In fact, such is the susceptibility of hamsters to
B. pseudomallei, that Miller et al. developed a method of purifying
B. pseudomallei from mixed cultures by infecting hamsters with the
contaminated sample and recovering the pure organism from the hearts blood
of the animals after they succumbed to the disease. 20
Strauss’ environmental surveys of the 1960’s used hundreds of hamsters to
purify samples. Filtered samples of soil or surface water were rendered isotonic
with saline then injected intraperitoneally into hamsters to yield pure cultures
of B. pseudomallei when they died five days later of melioidosis. 140 Hamsters
are not susceptible to B. thailandensis, 37 and it is interesting to speculate
whether B. thailandensis would have been discovered had hamsters continued
to be used in environmental surveys.
1.2.3 Melioidosis is an environmental organism
In 1955, Chambon demonstrated that B. pseudomallei could be isolated
from soil and water in South Vietnam (including rice fields), that the isolated
organism was virulent, and that it produced the characteristic histological
features of melioidosis in guinea pigs, rabbits and mice. 141 These findings were
confirmed and extended by a series of detailed epidemiological studies
published in 1969 by Strauss. 140,156,157 Strauss was able to culture
B. pseudomallei from 14·6% of the samples taken from rice paddy fields,
whereas samples taken from land used for all other purposes had isolation rates

The distribution of B. pseudomallei in soil is extremely irregular, and a site
with no detectable bacteria may lie as little as five metres from a site with
>10,000 cfu/g soil. 161 The reasons for this are unknown, but competition with
other Burkholderia species has been suggested as a reason. 162 Kaestli found that
in sites undisturbed by human activity, B. pseudomallei was more commonly
found in areas rich in grasses. In environmentally disturbed sites,
B. pseudomallei was associated with the presence of livestock animals and lower
soil pH. 163 It has been suggested that water-logged, clay soils may be better at
harbouring the bacterium than sandy, well-drained soils. 47,164 In common with
other Burkholderia species, 165 B. pseudomallei is able to survive and replicate
intracellularly within Acanthamoeba species in vitro. 118,166
Region of Thailand No. of positive sites Infection rate per
16
100,000 inpatients
Central 24·5% 13·4
North 13·8% 18
Northeast 50·1% 137·9
South 18·4% 14·4
Table 2. Isolation rate of B. pseudomallei from soil compared to
incidence of melioidosis by region of Thailand.
Adapted from Vuddhakul et al. Am J Trop Med Hyg 1999;60:458–61.
1.2.4 Melioidosis and rainfall
The association between melioidosis and rainfall was first noted by Strauss
in 1969, who found that he was unable to isolate B. pseudomallei from his
environmental samples when the weather was dry, but that samples became
positive after each rainfall. 140 Studies of human melioidosis cases in Australia,
Thailand and Singapore show a clear association with heavy rainfall in the 7
days preceding onset of the illness or the 14 days preceding admission to
hospital, and there is a positive correlation year-by-year between the amount of
rain and the number of cases of melioidosis. 167–171 There is certainly a
correlation with extreme weather events such as hurricanes. 172,173

Figure 5. The relationship between rainfall and melioidosis cases in
Ubon Ratchathani, Thailand.
Note.—This graph shows the quarterly incidence of melioidosis (solid line) over five
years (1987 to 1991), with the rainfall (interrupted line) in each quarter superimposed.
Figure from Suputtamongkol et al. Int J Epidemiol 1994;23:1082–90.
1.2.5 Human factors
B. pseudomallei is an organism of soil and surface water and a history of
contact with soil or surface water is therefore common, and a history of
residence in or travel to an endemic area is almost invariable. 169,174–177 In
Northeast Thailand, rice farming is an important part of the local economy. Rice
fields are flooded for much of the growing season and farmers spend much time
bare foot in paddy fields. Rice farming is therefore an important risk factor for
melioidosis. 169
Patients with melioidosis are more likely to be male. The ratio of males to
females is 1·41:1 in Thailand, 170 2·21:1 in Australia, 176 and 2·8–7·2:1 in
Singapore. 168 The statistics in Singapore are skewed by the fact that only males
are called up for military service, which involves high levels of exposure to
contaminated soil, and adjusting for this fact reduces the sex ratio to 1·3:1. 174
Harries et al. reported five cases of melioidosis in West African troops
(numbering approximately 30,000 in total) stationed in Burma, 129 and
17

marvelled that no cases were reported among Indian and Burmese troops
stationed at the same place. In Australia, aboriginals are overrepresented (24%
of the population, but 50% of melioidosis cases). 177 In Singapore, Indians and
Malays are at greater risk of melioidosis than the Chinese, 174,178 possibly related
to higher rates of diabetes in those two races. 179
Figure 6. Farming barefoot in northeast Thailand.
Note.—Farmers in Northeast Thailand walk through flooded paddy fields barefoot for
practical reasons. Images courtesy of Andrew Taylor © 2008 (left) and Vanaporn
Wuthiekanun © 2006 (right).
Diabetes mellitus is the most frequently identified risk factor in patients
with melioidosis: 39% of patients in Australia, 176 60% of patients in Thailand 175
and 76% in India have diabetes. 135 In Northeast Thailand, patients with diabetes
have 21·4-times greater risk of having melioidosis compared to people without
diabetes. 170 Heavy alcohol use is a risk factor in 12–39%, renal disease (chronic
renal failure or kidney stones) in 12–20·1%. Glucocorticoids are present in many
herbal preparations (called variously ya tom, yam mor or ya chut) which
patients use to self-medicate, and use of herbal medicines is an important cause
of immunosuppression in northeast Thailand. Other important risk factors
include thalassaemia and malignancy. 135,170,175,176 Patients with cystic fibrosis
may be particularly susceptibility; fortuitously, the endemic ranges for the two
diseases do not overlap, except in northern Australia. 52,137 HIV does not appear
to be a risk factor. 180
18

1.2.6 Routes of acquisition
The route of entry into the body is unknown. In Thailand, rice farmers
wade knee-deep in the muddy water of rice paddy fields, and rice farming is an
important risk factor for melioidosis. 175 By contrast, melioidosis is not a large
problem in rice-farming areas of Malaysia, where the process of farming has
been largely mechanised, and the direct exposure of farmers to soil and surface
water is limited. It has been suggested that the organism may gain entry via
scratches and small skin-breaks in the feet and legs, and there is one reported
case of inoculation following a snake bite. 40,125 In this hypothetical scenario,
pneumonia is explained by haematogenous spread. Against this hypothesis is
the fact that the majority of cases in Singapore (whose population is entirely
urban) have no history of contact with soil or surface water, despite a clear
association with rainfall. 168
The evidence supporting an inhalational route for melioidosis is largely
circumstantial. There is a clear relationship between rainfall and melioidosis
pneumonia, 167,169,170 and during the Vietnam War, an epidemiological link was
found between the helicopter crewmen and melioidosis infection, 13 of whom
many developed melioidosis pneumonia. This led to the conjecture that the
infection may have been due to inhalation of dust generated by the helicopter
rotors and that inhalation of aerosolized bacteria may be an important mode of
infection. 17,181 Data from animal models also appear to support this: the size of
inoculum needed to produce an equivalent effect in mice is four orders of
magnitude smaller than that needed to infect the same inbred strains
subcutaneously. 182 There is one case of laboratory-acquired melioidosis
reported in the literature due to inhalation of aerosolized bacteria, 183 so human
infection by this route is possible and viable bacteria have been found to be
aerosolized by severe weather events such as typhoons. 184
A third possible route of infection is via ingestion. Melioidosis has been
transmitted from mother to child via infected breast milk 185 and it is possible to
infect mice by oral gavage, but the doses required are of the order of 10 8 cfu, 186
which is an order of magnitude higher than that required for subcutaneous
infection, and at least six orders of magnitude higher than that required for
inhalation. Much of the population in Northeast Thailand draw their water from
wells or other untreated water-sources that may become contaminated with soil
19

during the rainy season. Two outbreaks in Northern Australia were associated
with contamination of drinking water, 187–189 but this does not necessarily mean
that the main mode of transmission in those outbreaks was ingestion, as potable
water was also used for washing and bathing.
The tsunami of Boxing Day, 2004, was associated with a spike in the
incidence of melioidosis in affected areas. 190–196 Cases of human-to-human
transmission have been described, but are exceptionally rare. 52,185,197,198 Vertical
transmission has been described but is extremely rare. 199
1.3 Clinical features of melioidosis
A definitive diagnosis of melioidosis may only be made on the basis of
microbiology (please refer to section 0 below), 40,41 and the main limitation to
diagnosis is the availability of adequate laboratory facilities.
Melioidosis has two main presentations: acute and chronic. Paediatric
disease is discussed in a separate section below.
1.3.1 Acute melioidosis
Acute melioidosis has no specific clinical features and should be suspected
in any septic patient with a history of residence in or travel to an endemic area.
The presence of a recognised risk factor (particularly diabetes) is helpful, but
the absence of risk factors does not exclude melioidosis. Where an inoculating
event has been noted, the mean incubation period is 9 days. 177
In Northeast Thailand, the median duration of symptoms prior to
presentation is 10 days. 200 Acute melioidosis is often disseminated, with
multiple sites being involved: 59–62% have bacteraemia and 49–56% have
pneumonia 200,201 (appearances on the chest radiograph range from lobar
pneumonia to widespread cotton wool patches); 28–52% have abscesses of the
spleen and/or liver 200–202 and 23–27% have skin and/or soft tissue infections
(principally wound infections and skin abscesses; cellulitis and erysipelas are
rare). 200,201 Other infections include those of the urinary tract (ranging from
pyelonephritis to prostatitis to asymptomatic bacteriuria), osteomyelitis, 203
septic arthritis, 203 myositis, 203 keratitis, 204 endophthalmitis, 205,206 orchitis, 207,208
and mycotic aneurysms. 209 Mortality is 42·6%. 170
Melioidosis in Australia presents in a similar manner to Thailand, with
some notable exceptions. Mortality is much lower (14%), which may be related
20

to differences in supportive care. 176 In Australia, 20% of male patients have
prostatic abscesses, 176 whereas this presentation is rare in Thailand. It is not
certain whether this is a true difference or whether it represents differences in
the availability of imaging (computed tomography of the pelvis is not routinely
performed in northeast Thailand).
There exists in Australia a syndrome of encephalomyelitis (2·6% of cases),
which manifests as a constellation of cerebellar signs, brainstem features and
cranial nerve palsies. 176 Few abnormalities are visible on CT, but dramatic
changes are visible on MRI. 210 By contrast, neurological melioidosis in
Southeast Asia (1·6% of cases in Northeast Thailand 211 ) presents with cerebral
abscesses (which may be multiple). These usually present with seizures and/or
focal neurological deficits and are clearly evident on CT. 212,211,213 In Singapore,
four of five cases of cerebral melioidosis were linked to sinusitis. 213 Primary
meningitis is extremely rare and there appears to be only one case reported in
the literature, 214 although meningitis has been reported to occur secondary to
rupture of a cerebral abscess. 40
1.3.2 Chronic melioidosis
Chronic melioidosis (defined as disease with onset ≥2 months prior to
presentation 215 ) comprises 12% of all adult melioidosis cases. 215 It usually
presents as localised disease, the majority of which are chronic non-healing
localised skin or soft tissue lesions, osteomyelitis, or pulmonary abscesses. It
often resembles tuberculosis clinically and is well described in the
literature. 26,216–220 Survival is the norm: of 30 cases of chronic melioidosis
reported in Northern Australia, all survived. 221
21

A
C
E
Figure 7. Clinical manifestations of melioidosis.
Note.—Pulmonary melioidosis has a wide range of presentations: A. Multiple bilateral
cotton wool patches. B. Left hilar lung abscess. C. Lobar pneumonia of the right upper
lobe. D. Large liver abscess in a 32-year-old female. E. Right renal abscess in a 63-year-
old female. F. Cutaneous abscess on the right arm of a 47-year-old male.
All images ©2007–2008 Gavin Koh.
D
22
B
F

The majority of early cases reported were acute and fulminant, but it was
recognised early on that in a small number of patients, the presentation could
have an indolent onset and chronic course. The first two cases of chronic
melioidosis were reported by Stanton & Fletcher in 1932, both with
osteomyelitis. Both patients survived to discharge, although one patient
continued to have discharging sinuses in his feet two years later. Grant and
Barwell describe a case of melioidosis in a 27-year-old soldier who present
initially in 1938 with septic arthritis of the right hip and ankle (initially
diagnosed as gonorrhoea), who went on to develop chronic pneumonia and
osteomyelitis of the frontal bone over the next five years. Mayer and Finlayson
describe the case of a 33-year-old soldier who developed bilateral lumbosacral
abscesses while stationed in Singapore in 1940. A year later, he developed a
chronic right lower lobe pneumonia and the lumbosacral abscesses were
recognised to have arisen from osteomyelitis of the lower thoracic spine. He was
initially diagnosed with tuberculosis, but the diagnosis was revised two years
after presentation in the light of B. pseudomallei being cultured from the pus.
The patient had been unwell for 4 years at the time the report was published. 216
1.3.3 Paediatric melioidosis
In Thailand, there are two age peaks: one in early childhood (≤10 years,
mean age 6·8 years) and one in middle age (≥45 years). 169,170,222 Paediatric
melioidosis in Thailand is largely a benign condition. Two-thirds of cases
present with localised disease 222 (38–40% of paediatric cases present with a
unilateral parotid abscess) and have a good prognosis. 223,224 A third present with
pneumonia or bacteraemia and have a presentation and prognosis similar to
adult disease. 222 Preliminary reports from Cambodia suggest that the spectrum
of paediatric disease there appears similar to that found in Thailand. 225
23

Figure 8. A 2-year-old child with unilateral parotitis at
Sappasithiprasong Hospital in Ubon Ratchathani, Thailand.
This 2-year-old male child developed swelling of his left cheek one week after playing in
a freshwater lake. The abscess was drained surgically and the child was discharged on
oral co-amoxiclav. Image © 2008 Gavin Koh. Signed consent obtained from parents.
In contrast to Thailand, Australia and Malaysia report no paediatric peak,
and when melioidosis does occur in children, the presentation and prognosis is
similar to when it occurs in adults. 215,226–228,227 Paediatric melioidosis in
Malaysia is associated with immunocompromise (particularly haematological
malignancy), 226 while paediatric melioidosis in Australia is associated with
cystic fibrosis. 227,229 Cultural differences may play a role: children in Northeast
Thailand are accustomed to playing in flooded paddy fields (splashing around
and catching the fish and crabs that grow there: both an important part of the
diet of Northeast Thailand). Their exposure to contaminated water is therefore
much greater than in elsewhere, and the syndrome of unilateral parotitis in
paediatric patients appears confined to Thailand. 215,226
1.3.4 Incubation period and latent melioidosis
The only systematic attempt to measure the incubation period in
melioidosis was made by Currie et al. in 2000: of 52 patients who reported an
24

inoculating event, the mean time to onset of symptoms was 9 days (range 1 to
21 days). 177
However, the incubation period for melioidosis may be much longer.
There were fewer than 300 cases of melioidosis among American soldiers
during the Vietnam war, but additional cases continued to present many years
after the war’s end, 230–232 earning it the name, ‘Vietnamese time bomb’. 233
Incubation periods of 18 years 234 and 28 years 235 have been reported in a
Vietnam veteran and a World War II veteran respectively. The longest reported
incubation period is 62 years, in an American man who had been taken prisoner
by the Japanese in 1942, who had worked building railroads in Singapore,
Malaysia, Burma and Thailand. After the war, he returned to Texas and did not
subsequently travel to a melioidosis endemic area. He presented in 2004 with a
non-healing ulcer on his right hand following a dog bite and wound culture grew
B. pseudomallei.
Latent melioidosis describes persistent clinically silent infection that may
activate many years after initial exposure or infection, and is named by analogy
to latent tuberculosis. We do not know how to diagnose latent melioidosis,
except retrospectively after reactivation. There is no information as to whether
serology for B. pseudomallei can identify latent melioidosis, or even whether
negative serology for B. pseudomallei excludes latent melioidosis.
Latent melioidosis seems to be uncommon and Currie et al. estimate that
latent disease accounts for no more than 4% of all cases. 177 There has not been
an explosion of melioidosis cases in America despite the fact that an estimated
225,000 soldiers returning from Vietnam were seropositive. 236 The
overwhelming majority of melioidosis cases are reported in endemic areas, are
of seasonal onset, and occur during the rainy season, all of which suggest that
recent acquisition and short incubation periods are the norm. 169,177 The available
evidence supports the conclusion that unlike latent tuberculosis, the majority of
patients with evidence of exposure to B. pseudomallei do not go on to develop
clinical melioidosis, even if they do have silent infection. The only appropriate
recommendation therefore is clinical vigilance.
Asymptomatic carriage of B. pseudomallei is rare. Currie et al. describe a
single case who was persistently sputum culture-positive for 12 months after
stopping antimicrobial chemotherapy. 221 A single case of asymptomatic throat
25

carriage in 58-year-old Vietnamese female rice farmer was reported by Phung et
al., 221,237,238 but Wuthiekanun found no instances of throat carriage on screening
3,524 Thai subjects. 239 Patients with cystic fibrosis may harbour
B. pseudomallei in their sputum even after appropriate treatment. 240,241
For all practical purposes, asymptomatic carriage of B. pseudomallei is
very rare. B. pseudomallei cannot be considered part of the normal human
flora, and the presence of even a single colony should be considered diagnostic
of melioidosis.
1.3.5 Seroprevalence
Melioidosis seroprevalence is defined as the prevalence in the general
population of people with a serum titre for anti-B. pseudomallei antibodies
within the diagnostic range for melioidosis.
As early as 1932, Dutch and English workers both dismissed serodiagnosis
as useless in the diagnosis of melioidosis, because titres in healthy controls were
so high. 11,242 Post World War II, Brygoo working in Vietnam (1953) and Nigg
working in Thailand (1963), both rediscovered the fact that healthy persons may
carry specific antibodies to B. pseudomallei in high titre. While evaluating
serology as a method for diagnosing melioidosis, Brygoo noted that 9·4% (24 of
255) Vietnamese controls had titres ≥1:40. 243 Nigg (apparently unaware of
Brygoo’s work) made her observations as part of a trial of a killed
B. pseudomallei vaccine conducted by the US military in Thailand. She found
that a proportion of the healthy Thai volunteers for her study had elevated titres
of antibody (as high as 1:5,120) even prior to immunisation.
B. pseudomallei is ubiquitously present in the soil in endemic areas, but
seroprevalence varies greatly and survey results are hard to compare because of
differences in study population (hospitalised patients versus community-based
studies) and in the cut-off titre selected (Table 3). The highest seroprevalence
recorded is in Northeast Thailand (Figure 9), 244 but seroprevalence throughout
Southeast Asia is high (Table 3). Positive serology is therefore extremely
common in otherwise healthy people in endemic areas such as Northeast
Thailand and this severely limits the utility of serological tests during the
investigation of suspected melioidosis.
26

Seroprevalence (cut-off) Study population Location Reference
7·3% (1:40) 1,592 adults Malaysia Strauss 1969 157
47·1% (1:40) 227 blood donors Thailand Khupulsup 1986 244
0·73% (1:16) 683 soldiers Singapore Yap 1991 245
83% (1:10), 22% (1:80) 295 children Thailand Charoenwong 1992 246
21% (1:40) 406 adults Vietnam Phung 1993 237
26·5% (1:40) 200 blood donors Malaysia Norazah 1996 247
8·7% (1:40) 160 adults Australia O’Brien 2004 248
38% (1:40), 7% (1:160) 968 adults Burma Wuthiekanun 2006 131
8·2% (1:40) 747 children Papua New Guinea Warner 2007 249
2·5% (1:40) 1,500 blood donors Australia Lazzaroni 2008 250
16·4% (any reaction),
6·5% (1:160)
968 children Cambodia Wuthiekanun 2008 251
Table 3. Summary of seroprevalence studies conducted using IHA.
Figure 9. Seroprevalence rises with age in Northeast Thailand.
The graph shows the distribution of IHA titres by age in 2,214 children from Northeast
Thailand. The proportion of children with any detectable titre rises rapidly in the first four
years of life then plateaus at 60–71%. By contrast, there is a gradual rise in the proportion
of children with a titre of ≥1:160, and by age 13, a fifth of children have an antibody titre
over 1:160. Figure taken from Wuthiekanun Am J Trop Med Hyg 2006;74:1074–5.
27

There is conflicting evidence whether the high seroprevalence seen in
endemic areas is due only to exposure to B. pseudomallei, or whether there is
cross-reactivity with B. thailandensis. Tiyawisutsri et al. examined 117 Thai
adult patients with culture-proven melioidosis and found that their sera did not
cross-react with B. thailandensis. 252 Gilmore et al. performed a similar study of
25 Australian patients with culture-proven melioidosis and found that their sera
did cross-react with B. thailandensis. 253 The methods used were similar except
that different strains were used in each study, but it is not clear whether this is
sufficient to explain the difference in results obtained.
1.4 Diagnosis
1.4.1 Microbiology
‘…there are, indeed, no cardinal symptoms upon which a trustworthy
diagnosis can be based, and the true nature of the disease can be determined
only by the cultivation of the causative organism…’
28
— A. T. Stanton & W. Fletcher, 1932
An early complaint (first made in 1932) was that melioidosis has no
cardinal symptoms or signs on which to base a diagnosis. 11 Indeed, the clinical
manifestations of melioidosis are protean and melioidosis most frequently
presents as undifferentiated sepsis. Infection has been described in almost every
organ in the human body, except that melioidosis seldom, if ever, causes
primary meningitis or endocarditis. 40
A definitive diagnosis of melioidosis is made by isolating B. pseudomallei
from any clinical specimen, because B. pseudomallei is not part of the normal
human flora. 40,41 Stanton & Fletcher’s 1932 monograph advised that a medical
officer suspecting melioidosis in a patient, but who had no access to a
laboratory, should inoculate guinea pigs with blood or pus, then send the
infected animals to the nearest laboratory for diagnosis. 11
The identification of even a single colony may clinch the diagnosis. Every
patient with suspected melioidosis should have a set of blood cultures, urine
specimen and a throat swab taken. Sputum, swabs from skin lesions, pus/fluid
from collections, etc. should be collected as indicated clinically. 41,239 Bone

marrow aspirates offer no benefits over blood culture. 254 Throat swabs may be
culture-positive regardless of the presence or absence of respiratory
symptoms. 239 Urine may be culture-positive even when there are no urinary
symptoms. Throat swabs and urine should therefore always be collected
regardless of symptoms or signs.
1.4.2 Serological diagnosis
Serological diagnosis is the identification of antibodies directed against
B. pseudomallei in the serum of blood from patients with melioidosis. Antibody
reactions to B. pseudomallei are common following environmental exposure,
are not protective, 250 and high titres may be found in patients with melioidosis.
Serological tests have therefore been the focus of the development of rapid
diagnostic tests for melioidosis.
The earliest reports measured the ability of patient serum to agglutinate
suspensions of whole bacteria. In 1932, De Moor et al., reporting from Batavia,
Dutch East Indies (modern day Jakarta, Indonesia), considered the use of
serology to diagnose melioidosis, but dismissed it as having little value, 242
because sera from normal controls agglutinated B. pseudomallei at titres
≥1:400. They therefore suggested that in the absence of positive cultures, only
titres ≥1:1000 could be considered diagnostic. Harries et al. (1948) suggested
that a titre of 1:80 be reported as ‘significant’ and that a titre of 1:160 was
‘diagnostic’, based on the observation that the majority of 56 control sera taken
from patients with other infections (principally syphilis) had no detectable titre
and none had titres >1:20. 129 The difference between de Moor’s and Harries’
results appears to be accounted for by the fact that the controls used by de Moor
were drawn from the native population (who would be expected to have a high
background exposure to B. pseudomallei), whereas the controls Harries used
were mainly British soldiers (who came from a non-endemic area and would
therefore not have had any exposure). Serological diagnosis based on
agglutinating whole bacteria continued to be reported sporadically until 1958. 25
Complement fixation tests have also been used. 255,256
The only validated serological test for melioidosis is the indirect
haemagglutination test (IHA). 257 IHA in its current form was described by
Strauss et al. in 1968 256 (although he was not the first to use IHA in melioidosis:
29

Nigg had described a version using human blood cells 5 years earlier 258 ). Sheep
erythrocytes are ‘sensitised’ with B. pseudomallei antigen derived from broth
cultures (called ‘melioidin’ by analogy to tuberculin) and the result reported is
the maximum dilution of patient serum that causes the sensitised erythrocytes
to agglutinate. The use of crude antigen in IHA means that the strain (or mix of
strains) used may result in differences in performance, and results from studies
performed using different strains may not be comparable.
Strauss collected sera from 1,592 healthy persons resident in Malaysia and
compared them to sera collected from 200 healthy persons resident in the US.
He set a 1:40 cut-off because none of the subjects in the US cohort had titres
greater than 1:20. He then showed that 7·3% of subjects in the Malaysian cohort
had titres ≥1:40. Alexander et al. validated the 1:40 cut-off for IHA on 169
melioidosis patients, using as controls, 201 US soldiers serving in Vietnam and
Thailand and 200 control sera collected in the US, 256 and concluded that the
1:40 cut-off was reliably specific for melioidosis.
A limitation of IHA clearly evident in Alexander’s study was the high
number of false negatives encountered in the first week of disease onset: only
38–45% were positive at the 1:40 cut-off during the first seven days. This is a
problem, because around half of all the patients who die of acute melioidosis die
within the first 48 hours of presentation. 199,200,259 In the study performed by
Alexander et al., the number of sera with reactions above the cut-off increased
to 88% in the second week, then plateaued at 89–100% in the third week.
Alexander was not the first to make this observation: de Moor reported this in
1932, and one of his arguments against the use of serodiagnosis was that
antibodies took too long to appear and serology was therefore useless in the
majority of patients. 242
This low sensitivity of IHA in acute melioidosis has been confirmed by
subsequent studies, with studies in Thailand tending to report slightly higher
sensitivities than elsewhere. Teparrugkul, in Ubon Ratchathani, reported a
sensitivity of 86% when using the 1:40 cut-off and 70% when the cut-off was
1:160. 260 In Australia, only 56% of patients with culture-confirmed melioidosis
had positive IHA results at admission (cut-off 1:40), 261 but most patients with
negative titres at admission subsequently seroconverted. An application for IHA
may therefore be in the diagnosis of patients with subacute or chronic disease
30

who are culture-negative, but in whom seroconversion or a rising titre can be
demonstrated.
The specificity of a test for melioidosis is defined as the percentage of
healthy people without melioidosis who test negative for melioidosis. The
specificity of IHA is therefore a function of the local seroprevalence for
melioidosis (speaking imprecisely, it is 100%–seroprevalence). Specificity of
IHA is therefore low in areas of high seroprevalence, because the number of
false positives is high.
Increasing the cut-off used increases the specificity of IHA at the expense
of decreasing the sensitivity. 262 In 1990, Appassakij et al. suggested a cut-off of
1:160 and evaluated this in 373 healthy blood donors, 65 cord blood samples
and 30 melioidosis patients. 263 He found that at this cut-off, sensitivity of the
test was 77% and specificity was 92%. Consistent with earlier studies, the test
was less likely to be positive in acute, fulminant melioidosis.
Not all studies from Thailand have been able to achieve such high
specificities. In the cohort reported by Teparrugkul (325 patients and 177
controls), a 1:40 cut-off gave a sensitivity of 86% and specificity of 38%; raising
the cut-off to 1:160 decreased sensitivity to 70%, but raised the specificity to
only 67%, 260 with similar results obtained from a later cohort at the same
centre. 264 In the cohort reported by Puapermpoonsiri (47 cases, 318 controls), a
1:40 cut-off gave a sensitivity of 83% and specificity of 90%, but increasing the
cut-off to 1:160 dropped the sensitivity to only 49% (specificity 97%). 265
The performance of IHA has been even poorer outside of Thailand. In
Malaysia, a cut-off of 1:40 for IHA resulted in a sensitivity of 36.0% and
specificity of 73·5%. 247 Raising the cut-off to 1:80 dropped the sensitivity to
29·0%, but increased the specificity to 80·5%. In Australia, Cheng et al. reported
a sensitivity of only 56% in 275 patients with culture-confirmed melioidosis. 261
An important exception is the island of Singapore, where melioidosis is
endemic but the background seroprevalence is low: only 0·73% of 683 men in
national service had IHA titres ≥1:16, sera from 50 healthy blood donors and 50
sewerage workers were all unreactive. 245 This means that cut-offs as low as 1:16
may be used without sacrificing specificity. The low seroprevalence seen in
Singapore may be explained by the fact that the population of Singapore is
31

100% urbanised. 266 Although the organism is present in the environment, the
majority of the population have little or no contact with soil or surface water.
The clinical significance of a positive IHA test in someone who is otherwise
well is not certain, but has been regarded as a marker for previous exposure to
B. pseudomallei. Its relationship to latent melioidosis is unknown. The
specificity of IHA for B. pseudomallei exposure has been disputed and there is
conflicting evidence whether exposure to the non-virulent bacterium,
B. thailandensis, might also be responsible for high IHA titres. 253,267,252
A number of other serological tests have been evaluated including an
indirect fluorescent antibody test for IgM specific antibody, 244 enzyme-linked
immunosorbant assays (ELISA) for B. pseudomallei lipopolysaccharide (LPS) 269
and for B. pseudomallei glycolipid, 270 as well as a commercial
immunochromogenic cassette (PanBio, Queensland, Australia). 264,248,270–272,262
None have so far supplanted IHA. 257
The role of serological methods has been revised in a recent study
comparing culture against four serological tests. Using Bayesian latent class
models, which allow for the fact that culture results are an imperfect gold
standard, the sensitivity of culture for diagnosing melioidosis has been
estimated to be only 60·2%. 274 By contrast, LPS-based ELISAs have reasonable
sensitivity (80·2%) and specificity (95·0%), and may be good candidates to take
forward for further development. 262
1.5 Therapy
1.5.1 Historical perspective
The majority of reports of melioidosis in the pre-antibiotic era were post
mortem. The mainstay of treatment was largely supportive, with drainage of
clinically evident collections and debridement of infected tissue. Non-
chemotherapeutic therapies tried include autogenous vaccination (isolating the
organism, culturing it in a guinea pig, and then repeatedly inoculating the
patient with the cultured bacterium). 275
The first reported attempt using modern antimicrobial chemotherapy to
treat melioidosis was in 1943, when Grant and Barwell attempted to treat
melioidosis osteomyelitis in a soldier with sulphadiazine. Although a good
response was achieved, the patient relapsed each time the treatment was
32

stopped. 26 Harries et al. noted that sulphonamide monotherapy was ineffective
in disseminated disease. 129
It was discovered early on that penicillin was ineffective against
B. pseudomallei in vitro, 275–277 as was the other wonder drug, streptomycin. 278
However, chloramphenicol 279–281 and the tetracyclines 281,282 were active. In
1952, Dunlop reported the successful treatment of a woman with melioidosis of
the ovary with surgical drainage and chlortetracycline. 283 However, any initial
enthusiasm for the tetracyclines was tempered by the realisation that resistance
develops quickly both in vitro and in vivo. 284
B. pseudomallei is intrinsically resistant to a wide variety of antimicrobial
agents, including all penicillins (except the anti-pseudomonal penicillins), 280,285
all macrolides, 286 all first and second-generation cephalosporins 280 and
polymyxins. 287 It is intrinsically resistant to all aminoglycosides, except that
some strains are sensitive to kanamycin in vitro. 288,77,280 However, the
aminoglycosides, as a class, have no intracellular penetration 286 and are
therefore ineffective clinically. There is one reported case of a woman who was
successfully treated with kanamycin, 183 but this was in combination with
tetracycline, to which the organism was susceptible.
Until 1989, in Thailand, the conventional treatment for melioidosis was
combination therapy with co-trimoxazole, doxycycline and chloramphenicol. 259
1.5.2 Current recommendations
In its most serious manifestation, melioidosis causes severe sepsis
associated with hypotension and multiple organ failure, requiring intensive
supportive therapy (fluid management, invasive monitoring and artificial
ventilation). These measures likely play as great a role in the treatment of
melioidosis as do antibiotics. 289 Human-to-human transmission is very rare and
patients may be safely accommodated on the open ward. 102-105
Antimicrobial therapy for melioidosis is divided into two phases: an
intensive phase (parenteral inpatient therapy) and an eradication phase (oral
outpatient therapy). 41 Current treatment recommendations are summarised in
Table 4.
33

1.5.3 Intensive phase
Antibiotic therapy should be initiated as soon as possible, as there is
evidence from other causes of sepsis that delay in initiating appropriate
antimicrobial therapy increases mortality. 290,291 Without access to antimicrobial
therapy, mortality from the bacteraemic form of melioidosis approaches 100%
and in the pre-antibiotic era, almost all case series reported in the literature
were post mortem. 127,9
Ceftazidime treatment was introduced in 1989, after it was shown that it
halved mortality compared to conventional therapy (co-trimoxazole,
doxycycline and chloramphenicol). 259 The first-line treatment for melioidosis is
high-dose intravenous ceftazidime 41,259 for ≥10 days, continued until there are
clear signs of clinical improvement (principally defervescence). The median
fever clearance time is 10 days despite adequate antimicrobial chemotherapy. 200
Failure to defervesce is not therefore itself an indication to change treatment
and it is not uncommon for patients to require over a month of intravenous
therapy.
Carbapenems (principally, meropenem and imipenem) are routinely used
in Australia. 132 Carbapenems may be preferred to ceftazidime because the
minimum inhibitory concentrations (MIC) for imipenem and meropenem
against B. pseudomallei are lower than those for ceftazidime, 288,292,77 and
carbapenem therapy is associated with lower levels of antibiotic-induced
endotoxin release. 293 However, a head-to-head comparison of imipenem versus
ceftazidime found no difference between the regimens, but the trial was stopped
early owing to withdrawal of drug company support. 200 A multi-centre double-
blind randomized-controlled trial comparing meropenem against ceftazidime
started recruiting in 2007 (ClinicalTrials.gov NCT00579956).
The addition of co-trimoxazole to standard ceftazidime therapy provides
no benefit, 294,295 despite evidence from animal models to support the practice. 296
However, co-trimoxazole has excellent tissue penetration into brain and
prostate, so there is a theoretical benefit to adjunctive co-trimoxazole when
treating these infections. 41 Co-amoxiclav is an alternative to ceftazidime if none
of the first-line drugs are available: 297 mortality is the same, but the risk of
treatment failure is higher. Co-amoxiclav may be useful in resource-poor
settings where first-line drugs are not immediately available.
34

High intensive parenteral a Typical dose b Duration
ceftazidime 40 mg/kg (max. 2 g) tds, c or
meropenem 25 mg/kg (max. 1 g) tds, or
imipenem 20 mg/kg (max. 1 g) tds
2 g every 8 h c
1 g every 8 h
1 g every 8 h
Alternatively,
co-amoxiclav 20/4 mg/kg (max. 1000/200 mg) qqh e 1000/200 mg every 4 h
Eradication oral phase (Thailand)
co-trimoxazole (60kg: 320/1600mg bd) +
doxycycline 2 mg/kg (max. 100 mg) bd
320/1600 mg every 12 h
100 mg every 12 h
35
≥10 days, or until clear clinical
improvement, which ever is longer d
12–20 weeks
Alternatively,
co-amoxiclav 60/15 mg/kg (max. 1500/375) tds f Eradication oral phase (Australia)
1500/375 mg every 8 h
co-trimoxazole (60kg: 320/1600mg bd)
320/1600 mg every 12 h 3–6 months
Table 4. Summary of antimicrobial chemotherapy of melioidosis.
Note.— bd = bis die; h = hours; max. = maximum; qds = quater die sumendus; qqh = quaque quarta hora; tds = ter die sumendus. The regimen
quoted for Thailand is that used at Sappasithiprasong Hospital, Ubon Ratchathani. The regimen quoted for Australia is that used at the Royal
Darwin Hospital, Northern Territory. All doses require adjustment for renal function.
a In Australia, oral co-trimoxazole is routinely added to parenteral therapy for infection of the central nervous system and for prostatitis, because of
the excellent penetration of co-trimoxazole to these sites.
b Typical dose for a 70 kg adult with normal renal function.
c The Royal Darwin Hospital uses 50 mg/kg (maximum 2 g) qds. A typical dose is therefore 2 g every 6 hours.
d Brain abscesses are routinely treated with ≥4 weeks of parenteral antibiotics. Other deep seated abscess that are not amenable to drainage may
also require prolonged parenteral therapy.

e There is no parenteral preparation of co-amoxiclav available in Australia. Ampicillin-sulbactam is available in a parenteral formulation, but there
is no clinical experience of the use of this drug in melioidosis.
f If co-amoxiclav preparations are only available in a 2:1 ratio, then additional amoxicillin should be prescribed to achieve the correct ratio of
amoxicillin to clavulanate. Co-amoxiclav is routinely used in children and pregnant women in Thailand, because doxycycline is relatively contra-
indicated.
36

Primary resistant to first-line agents is unusual (Table 5) and there is no
evidence that these rates are changing. B. pseudomallei is an environmental
organism, with no human-to-human transmission, and is largely unaffected
therefore by antibiotic pressure in healthcare settings. Rarely, resistance can arise
during the course of treatment, 298,299 and patients who remain febrile on treatment
should have cultures repeated regularly to look for this. 299
Cefoperazone-sulbactam has been shown to be equivalent to ceftazidime in a
randomised controlled trial. 201 There are no obvious circumstances under which
cefoperazone-sulbactam would be preferred to ceftazidime or a carbapenem, and
the drug has not found a home in current guidelines. There is in vitro evidence for
biapenem, 292 doripenem, 300 piperacillin-tazobactam 184,301 and tigecycline, 302–304
but no clinical experience to support their use.
CAZ IPM MEM AMC Region Reference
99.5% 100% – 100% Thailand Dance 1989
99% 100% 100% 100% Thailand Smith 1996 292
100% 100% 100% 99.4% Australia Jenney 2001
100% 100% 100% 100% Hong Kong Ho 2002 184
98% a 100% a – 98% a Multiple Thibault 2004 77
– – – 100% Singapore Sivalingam 2006 305
– 100% 100% – Malaysia Karunakaran 2007 306
99·4% 99·4% – 97·9% Singapore Tan 2008 285
Table 5. Susceptibility of B. pseudomallei to first-line parenteral
agents.
Note.—CAZ = ceftazidime; IPM = imipenem; MEM = meropenem; AMC = co-amoxiclav.
These results are for primary isolates and do not include data from resistance acquired
during the course of treatment. A dash indicates that that agent was not tested. The
breakpoints used are those defined by the Clinical Laboratory Standards Institute in the US
(CAZ ≤8 mg/l; IMP ≤4; MEM ≤4; AMC ≤8/2).
a Breakpoints defined by Comité de l’Antibiogramme de la Société Française de
Microbiologie, France (CAZ ≤32 mg/l; IMP ≤8; AMC ≤16/2).
37

1.5.4 Eradication phase
Thirteen per cent of patients who survive a first episode of melioidosis go on
to experience a recurrence of their disease: 221,307 75% of these are due to relapse
with the same strain and 25% are due to re-infection. 307 The aim of eradication
therapy is to reduce the risk of relapse.
Eradication therapy is based on oral medication and is usually conducted on
an outpatient basis. Patients are not started on oral eradication therapy until they
are clinically well and ready for discharge home. The occasional patient with a
solitary abscess that is easily drained, who is afebrile, and who has no systemic
signs or symptoms, may avoid the intensive phase and go straight onto oral
eradication therapy.
Relapse is more likely in patients who have multifocal disease or bacteraemia
at their first presentation. 307 Regimens containing co-trimoxazole have the lowest
relapse rate. Ciprofloxacin monotherapy, ciprofloxacin/azithromycin combination
therapy and doxycycline monotherapy are associated with higher relapse rates and
cannot be recommended. 307–310
Unfortunately, none of the first-line treatments (ceftazidime, meropenem,
imipenem) may be administered orally as they are not absorbed.
Standard therapy in Thailand is combination therapy with trimethoprim-
sulphamethoxazole and doxycycline for 12–20 weeks; 311 standard therapy in
Australia is co-trimoxazole alone. 132 It is not clear that one therapy is superior to
the other, and a double-blind randomised placebo-controlled trial to compare the
two regimens, MERTH (Melioidosis Eradication THerapy/THailand) has just been
completed. The study finished recruitment in 2009 and completed follow-up in
2010 (trial registration ISRCTN86140460). A report is expected at the end of
2011. 312
One potential benefit of co-trimoxazole monotherapy is that co-trimoxazole
and doxycycline are antagonistic in vitro. 313 It is also anticipated that a simplified
regimen would be better tolerated, because those effects due to doxycycline would
be eliminated (specifically, oesophageal ulceration and photosensitivity, but also
non-specific gastrointestinal effects such as nausea and vomiting) and because of
the smaller pill burden. The dose of co-trimoxazole used in Thailand was revised
upwards in 2008 on the basis of evidence from pharmacokinetic modelling to be in
line with that used in Australia. 314
38

1.5.5 Vaccination
There are currently no licensed vaccines for melioidosis. Currently, the best
vaccine candidates are live attenuated vaccines, but none of these provide
sterilising immunity and there are no vaccines currently in clinical trials. 315,316
1.5.6 Post-exposure prophylaxis
Most cases of accidental exposure are likely to occur in a clinical laboratory
processing a specimen not known or suspected to have B. pseudomallei. Given the
lack of other evidence, recommendations for post-exposure prophylaxis are based
on current treatment recommendations. Following a thorough risk assessment,
post exposure prophylaxis with three weeks’ co-trimoxazole may be offered. 317
Workers intolerant of co-trimoxazole may be offered co-amoxiclav instead.
Workers who seroconvert may be offered an extended course of therapy (12 weeks)
and careful monitoring for the development of clinical disease.
There are not currently any published recommendations for travellers with
risk factors such as diabetes or cystic fibrosis who travel to endemic areas.
1.5.7 Antisepsis
B. pseudomallei is susceptible to 70% ethanol, 1% hypochlorite, 7% tincture
of iodine, 0·05% benzalkonium chloride and 1% potassium permanganate. 20,318 The
commercial agents, Perasafe (DuPont, active ingredient peracetic acid) and
Virkon® (DuPont, active ingredient commercial confidential, but known to
include peracetic acid) are also bactericidal for B. pseudomallei. 318
Five per cent phenol does not reliably disinfect B. pseudomallei. 20 Outbreaks
of melioidosis have been associated with contaminated hand wash. 319,320
Chlorination effectively treats drinking water: free chlorine concentrations
normally used by water purification plants in the West are ~1 mg/l, which is
sufficient for the majority of strains of B. pseudomallei, but some Australian
strains require ≥30 mg/l. 166,321–323
1.6 Adjunctive therapies
1.6.1 Granulocyte-colony stimulating factor
The only adjunctive treatment ever studied in the specific context of
melioidosis was granulocyte colony stimulating factor (GCSF). 324 GCSF is a 20 kDa
glycoprotein growth factor produced principally by endothelium and macrophages,
39

that stimulates the proliferation and differentiation of granulocyte precursors in
the bone marrow. 325 Its principal medical application is to reduce the duration of
neutropenia that occurs in cancer chemotherapy and HIV. As a drug, its generic
name is filgrastim, and is sold in the UK under the trade names Neupogen®,
Ratiograstim® and Zarzio®.
The neutrophil response is critically important in the pathogenesis of
melioidosis, 326 and the introduction of adjunctive therapy with GCSF coincided
with a reduction in melioidosis mortality from 95% to 10% at one Australian
centre. 327 Unfortunately, a single-centre randomised-controlled trial found no
evidence of benefit. 324
1.6.2 Intensive glucose management
On the basis that plasma glucose levels are often high in sepsis and that high
levels correlate with mortality, it has been proposed that intensive insulin therapy
to tightly control glucose levels in the intensive care setting should decrease
mortality in critically ill patients (not just patients with sepsis). 328 In a study of
1,548 surgical intensive care patients in Belgium, intensive insulin therapy (target
range 4·4–6·1 mM) reduced mortality from 8·0% to 4·6%. 221 This is potentially
relevant, because 57% of patients with melioidosis have diabetes or present with
hyperglycaemia, 329 and, insulin/ glucose monitoring are routinely available at
provincial hospitals in Thailand.
The van den Berghe study 221 was criticised on a number of counts:
principally, that this was a single centre study, most of the patients were admitted
post-surgery, the study was open label, and the feeding regimen was unusual (all
patients received a 200–300 g bolus of glucose at recruitment and total parenteral
nutrition or enteral feeding was started within 24 hour of admission). 330 A repeat
study by the same investigators in the medical intensive care was unable to
replicate this finding. 331 Studies of intensive glucose control in the context of
myocardial infarction 332 and trauma 333 have also found no benefit. Conversely, a
multi-centre trial of 6104 patients by other investigators found that intensive
insulin therapy (target 4·5–6·0 mM) in critically ill patients increased mortality. 334
A meta-analysis of 26 trials (13,567 patients, including the data from NICE-
SUGAR) 335 found that intensive insulin therapy increases the risk of
hypoglycaemia without conferring any over mortality benefit to critically ill
40

patients, and the only subgroup of patients found to benefit were surgical patients
(pooled relative risk 0·63, 95%CI 0·44–0·91). 335
1.6.3 Immunosuppressants
A recurring theme in the sepsis literature is that the inflammatory response is
essential, yet high levels of inflammation correlate with mortality in observational
studies of human sepsis. 336 The prevailing theory has been that mortality from
sepsis is due to an uncontrolled inflammatory response: 336 the ‘cytokine storm’. 337
This view has been supported by data from animal models of sepsis, 338–344 and that
has in turn driven the clinical development of anti-inflammatory drugs as
adjunctive therapies in sepsis.
Multiple trials looking for a benefit from high-dose corticosteroids in severe
sepsis and septic shock have not found any benefit, with some studies finding a
detrimental effect. 345–348 Evidence for the use of low-dose corticosteroids in sepsis
is mixed: a single multi-centre, placebo-controlled, randomised, double-blind trial
of 300 intensive care patients with septic shock and relative adrenal insufficiency
(as measured by a cortocotrophin test) found a mortality benefit from low-dose
corticosteroid and fludrocortisone, 349 but this result has not been possible to
replicate: one trial found that low-dose corticosteroid reduced time on
vasopressors but not mortality, 350 another multi-centre randomised placebo-
controlled trial of 248 patients performed by the same investigators found no
difference in mortality 351 and a further trial in patients with liver cirrhosis found
that the patients in the hydrocortisone arm had an increased incidence in relapse
of shock. 352
The results for anti-tumour necrosis factor-alpha (TNFα) therapy have been
disappointing. Despite evidence from primate models that anti-TNFα has a
dramatic effect on survival from Escherichia coli 353,354,338 and Staphylococcus
aureus bacteraemia, 339 the result from multiple clinical trials is that anti-TNFα
therapy provides no benefit in sepsis. 355–359 In any case, results from mouse
models of melioidosis would predict TNFα blockade to be detrimental. 360 Trials of
interleukin-1β beta blockade have been similarly disappointing, 361–363 despite
promising results from animal models. 340–344
41

1.6.4 Activated protein C
Activated protein C (APC) is a serine protease that is known primarily as an
anticoagulant. Inactive protein C circulates in the plasma and is cleaved to its
active form by activated thrombin. As a drug, its generic name is drotrecogin alfa
(activated), and its trade name is Xigris®.
APC is the first (and currently, the only) US Food and Drug Administration
(FDA)-approved adjunctive treatment for severe sepsis. 364,365 Its approval was
based on the results of a Phase III multi-centre double-blind placebo-controlled
trial involving 1,690 sepsis patients in 164 centres in 11 countries (the PROWESS
trial). 366 The trial results were published in 2001 and reported a 6% reduction in
the absolute risk of death, accompanied by a 1·5% increase in the risk of serious
bleeding events.
Levels of protein C are low in melioidosis, 367,368 and we may speculate that
melioidosis patients with severe sepsis may benefit from APC therapy. No trials of
activated protein C therapy are currently planned in melioidosis.
1.6.5 Gamma interferon
Gamma interferon (IFNγ) is necessary for the normal host defence against
B. pseudomallei infection and IFNγ levels are indeed elevated in acute
melioidosis. 369 Small doses of IFNγ greatly potentiate the effect of ceftazidime in
vitro and mouse models. 370 IFNγ supplementation is already used in the
management of multi-drug resistant tuberculosis, and it seems reasonable to
hypothesise that patients with other infections caused by intracellular bacteria may
benefit. 371 Using a mouse model of melioidosis, Propst et al. found that IFNγ
supplementation reduced 20-day mortality in ceftazidime-treated animals from
100% to 30% (p

a model organism do not necessarily translate to the clinical situation, and findings
made in one model may not necessarily correlate with findings made in another
model. Interpretation of the literature is therefore an exercise akin to joining dots
to form a picture, except that the dots are unnumbered, some are optional, and
many are not even printed on the same page.
A B
Figure 10. Histological features of human melioidosis.
Note.—A Lung section taken post mortem from an 18-year-old female. Tangles of
extracellular bacteria are indicated by large arrow heads. Tangles of intracellular bacteria
within macrophages and giant cells are indicated by thin arrows (magnification ×1000). B A
splenic granuloma with central necrosis (magnification ×100) surrounded by foamy
macrophages and multinucleate giant cells (inset, magnification ×400). Figures adapted from
Wong Histopathology 1995;26:51.
1.7.1 Histology
There are few histological descriptions of human melioidosis from
Thailand because of strong cultural barriers to the collection of tissue
specimens. The most recent descriptions come from a series of five post mortem
cases and 14 biopsy specimens from Malaysia, 52 although older descriptions
exist from the Vietnam War 17,374,375 and from before World War II. 11
43

In the Malaysian series, Wong et al. describe areas of necrotising
inflammation containing mixtures of neutrophils, lymphocytes, macrophages and
multinucleate giant cells. Numerous bacteria are seen within these lesions and may
occur both intra- and extracellularly (Figure 10). When they occur intracellularly in
macrophages, the bacteria may be so numerous as to form globi. The lesions may
sometimes form granulomas with central necrosis and surrounding foamy
macrophages, which cause them to resemble tuberculosis or cat-scratch fever.
1.7.2 The gamma interferon response in melioidosis
B. pseudomallei is a facultative intracellular pathogen, 79–81 and the host
cytokine response to melioidosis may be appropriately summarised as an IFNγ
response to an intracellular pathogen. 369,376,377 While IFNγ elevation is not a
general feature of other causes of sepsis, 378,379 it is a feature of infection caused by
other intracellular organisms, such as Brucella species, 380 Mycobaterium
tuberculosis, 381,382 Neisseria meningitidis, 383–385 and Salmonella Typhi. 386 This is
to be expected, given that a primary function of IFNγ is to stimulate the killing of
intracellular pathogens. 387–389
The interferons were initially discovered in 1947 by Henle and Henle as a
substance produced in the allantois of chick embryos infected with influenza that
‘interfered’ with viral replication. 390 This substance was given the name interferon
by Isaacs and Lindeman in 1957, 391 but in the mid-1970’s, it was realised that
interferon was not one protein but a mixture of proteins. 392,393 In 1979, Grob and
Chadha separated human ‘interferon’ into three fractions, later designated alpha,
beta and gamma. 394 Macrophage activating factor was first isolated in 1973 from
guinea pig lymphocytes and found to activate the intracellular killing effects of
macrophages, 395 but it was cloning of the IFNγ gene (IFNG) in 1982 396,397 that
brought the realisation that IFNγ and MAF are the same molecule. 398–402
In 1991, Brown et al. studied 65 melioidosis patients and found that serum
concentrations of IFNγ at admission were elevated in melioidosis, and that the
higher the serum concentration of IFNγ, the more likely the patient was to die. 369
There are two possible interpretations for this finding: the first is that mortality is
due to an inappropriate and excessive inflammatory response. 403 The second is
that the inflammatory response is appropriately high, that it correlates with higher
bacterial loads, disease severity, and hence, mortality. 376 The distinction is an
44

important one: if the first is true, then suppressing IFNγ would be expected to
decrease mortality; if the second is true, then suppressing IFNγ would be expected
to increase mortality. Similar arguments may be made for other cytokines.
Evidence for the first hypothesis comes from murine models of endotoxic
shock (a model of septic shock induced with intravenous LPS). 404,405 Blocking
antibodies directed against IFNγ prevented mortality from endotoxic shock 404 and
IFNγ receptor knockout mice were also able to tolerate 100 to 1000-times more
LPS than wild type mice. 405 However, the clinical relevance of data obtained from
endotoxic shock models has been questioned repeatedly. 406–409
Evidence in favour of the second hypothesis was obtained by Lauw et al. in
her survey of 62 consecutive melioidosis patients. 376,377 Although she did not
quantify bacterial loads, she found that bacteraemic patients had higher levels of
IFNγ. 376 In murine models of melioidosis, blocking IFNγ increased susceptibility in
Taylor outbred mice, lowering the intraperitoneal LD50 from 5 × 10 5 cfu to
2 cfu, 373 whereas exogenous IFNγ supplementation improved survival. 370 Taken
together, the experimental and clinical evidence support the conclusion that IFNγ
levels are appropriately elevated in melioidosis in response to the intracellular
nature of the pathogen and the severity of the disease.
In mice, the primary source of the IFNγ are natural killer (NK) cells and
memory CD8 + T-lymphocytes, 360,410 and this IFNγ response appears to be
secondary to the synergistic action of IL12 and IL18. 360,410 Ablating either IL12 or
IL18 ablates the IFNγ response and worsens mortality. 360,411,412 This is supported
by the clinical data, which shows elevations of IL18 in plasma 376 as well as
elevations of IL18 mRNA in blood mononuclear cells 412 in melioidosis patients.
45

Cytokine Primary source(s) Summary of functions References
CXCL9 macrophage,
various
CXCL10 macrophage,
various
IFNγ NK, CD8,
(macrophage)
Secreted in response to
IFNγ stimulation.
Chemoattractant for T cells
and NK cells.
Secreted in response to
IFNγ stimulation.
Chemoattractant for T cells
and NK cells.
Secreted in response to IL12
and IL18, Stimulate
macrophages and killing of
intracellular bacteria.
IL6 macrophage Downstream of TLR
activation. Stimulates
monocyte differentiation to
macrophage. Inhibits IL1β
IL8 macrophages,
endothelium
IL12 monocytes,
macropha
ges 414,415
IL15 monocytes,
macrophages
IL18 monocytes,
macrophages
and TNFα.
Downstream of TLR
activation. Recruits
neutrophils.
Downstream of TLR
activation. Stimulate IFNγ
production. 416
Stimulates proliferation of
NK cells. 416
Produced by inflammasome
activation. Stimulate IFNγ
production. 360,411
TNFα macrophages Downstream of TLR
pathways. Amplifies IFNγ
production.
46
Lauw 2000 377
Lauw 2000 377
Santinarand 1999, 373
Lauw 1999 376
Friedland 1992 413
Friedland 1992 413
Lertmemongkolchai
2001, 360 Koo 2006 410
Lauw 1999 376
Lertmemongkolchai
2001, 360 Koo 2006 410
Santinarand 1999, 373
West 2008 417
Table 6. Cytokines and chemokines involved in clinical melioidosis.
Note.—CD = cluster of differentiation; CD8 = CD8-positive T-lymphocytes; IFN = interferon;
IL = interleukin; NK = natural killer cells; TLR = Toll-like receptor; TNF = tumour necrosis
factor. This table is restricted to those cytokines and chemokines that have been measured in
clinical melioidosis. This table is not intended to be a comprehensive list of cytokines
known to play a role in the host response to melioidosis.

TNFα levels were measured in a clinical study of 55 melioidosis patients,
and the presence of detectable plasma TNFα at admission correlated with
mortality. 418 In mice, ablating TNFα reduces the magnitude of the IFNγ
response by up to 30%, but does not ablate it. 360 The relationship of TNFα to
IFNγ is circular, with TNFα being both upstream 373 as well as downstream of
IFNγ secretion. 419,420 One interpretation consistent with these data is that the
role of TNFα in melioidosis is to amplify the IFNγ response, but not to initiate it.
Downstream of IFNγ, CXCL9 (formerly called MIG) and CXCL10 (formerly
IP-10) are secreted by a large variety of cell types in response to IFNγ
stimulation. Plasma concentrations of both CXCL9 and CXCL10 are elevated in
patients with melioidosis and levels were higher in patients with bacteraemia. 377
CXCL9 and CXCL10 are T-cell and NK cell chemoattractants 421 and probably
work to amplify the IFNγ response.
In clinical studies of melioidosis, plasma concentrations of IL6, IL8 and
IL10 have also been found to be elevated and to correlate with mortality (Table
6). 413,422
1.7.3 Pattern recognition receptors
Pattern recognition receptors (PPRs) are a set of evolutionarily-conserved
receptors present on leukocytes to identify pathogen-associated molecular
patterns (PAMPs), e.g., the lipopolysaccharide of Gram-negative bacteria,
lipoteichoic acid of Gram-positive bacteria and viral RNA. In evolutionary
terms, these receptors predate the development of adaptive immunity and allow
the innate host response to identify antigens that are not ‘self’.
The best studied of the pattern recognition receptors are the Toll-like
receptors (TLRs) and the Nod-like receptors (NLRs).
1.7.4 Toll-like receptors and nuclear factor kappa B
There are 13 human and mouse Toll-like receptors (TLR) described to
date. 423–425 The TLRs are membrane-bound glycoprotein receptors and are
members of the interleukin-1 receptor superfamily. 424 The TLRs are one of the
more evolutionarily conserved parts of the innate immune system and may have
originated in the common ancestor of all bilaterian animals. 426,427 TLRs
signalling (with the exception of TLR3) proceeds via the adaptor molecule,
47

MyD88, 424 and the importance of the TLRs in the host response to
B. pseudomallei infection has been demonstrated by the fact that Myd88
knockout mice are extremely susceptible to B. pseudomallei infection. 428 Mouse
gene expression studies have also found upregulation of Tlr2, Tlr4 and Tlr7 in
experimental melioidosis. 429
A major downstream effect of TLR signalling is nuclear factor kappa B
(NF-κB) activation. NF-κB is a family of constitutively expressed protein
complexes that function as DNA transcription factors. These complexes are
homo- or heterodimers of a family of five proteins, RELA (also called p65),
RELB, c-REL, NF-κB1 (also called p50 and p105) and NF-κB2 (also called p52
and p100).
NF-κB is activated by a wide range of stimuli, including cytokines (notably
TNFα and IL1 430 ), growth factors, mitogens, oxidative stress, chemical stress,
bacterial and viral antigens. Activation of NF-κB causes it to move from the
cytoplasm to the nucleus, where it regulates the expression of several hundred
genes, including several cytokines and chemokines (e.g., IL2, IL8, IL12, GM-
CSF and CD40 ligand). 431,432 IL8 secretion is often used as a marker for TLR
activation, because it is produced in very large quantities. 152 NF-κB is a rapidly
acting transcription factor, due to the fact that it is constitutively expressed in
the cytoplasm of cells and does not require new protein synthesis for its
activation.
Aside from cytokine secretion, NF-κB activation also promotes cell
survival and proliferation. 433,434 Although TLR activation was initially believed
to result in a non-specific NF-κB response, it is now known that specific
responses do occur: for example, TLR4 and 9 activation produces a Th1-type
response, but TLR2 activation results in T regulatory cell activation. 435 The
picture is also complicated by the fact that there also exists extensive crosstalk
both within the TLR family and between TLRs and other PRRs. 436,437
The best described TLRs are members 1 to 9. Of these, TLRs 1, 2, 4, 5 and
6 are located on the cell membrane where they form dimers. TLR2/TLR1 and
TLR2/TLR6 form heterodimers that recognises triacyl lipopeptide and diacyl
lipopeptide respectively, and both will recognise lipoteichoic acid. 427,438 TLR4
and TLR5 form heterodimers that classically recognise lipopolysaccharide and
48

flagellin respectively. 427 TLRs 3, 7, 8 and 9 are found intracellularly on the
endosomal membrane, where they recognise single and double stranded RNA,
and unmethylated CpG oligodeoxynucleotides. 427
In clinical melioidosis, Wiersinga et al. found increased expression of
TLR1, TLR2, TLR4, TLR5, and TLR8 in circulating blood leukocytes, 152 while
levels of TLR3 and TLR7 were no different from controls. The lack of differential
expression of TLR3 is unsurprising, as this TLR is restricted to dendritic cells,
which do not circulate; this is reflected in the low absolute expression values
reported. 439 TLR7 is constitutively expressed by human plasmacytoid dendritic
cells (PDC) and B lymphocytes: upon stimulation, expression of TLR7 increases
in PDCs but decreases in B lymphocytes. 440 Wiersinga’s study did not
distinguish between types of mononuclear cells and was therefore not designed
to detect this type of differential expression.
The increased expression of TLR4 accords with the fact that
B. pseudomallei is a Gram-negative organism and, as such, expresses the
classical TLR4 ligand, LPS. 99,441,442 Other proteins in the endotoxin receptor
complex, CD14 and LY96 (=MD2), were also upregulated in melioidosis
patients 443 and it has also been reported that TLR4 polymorphisms are
associated with an increased susceptibility to melioidosis, 444 however, the
evidence for a role of TLR4 in melioidosis is contradictory. TLR4 appears to
have no role in the innate host response to B. pseudomallei as evidenced by the
fact that Tlr4-knockout mice had no phenotype when compared to wild-type
mice on a C57Bl/6 background. 152 Supporting evidence was provided by two
studies using the closely-related organism, B. thailandensis, in Tlr4-knockout
mice as well as the Tlr4-deficient C3H/HeJ mouse strain, both of which showed
that TLR4 is not a major player in sepsis caused by B. pseudomallei. 445,446 A
possible explanation for the lack of a TLR4 response may be due to structural
differences in the lipid A moiety of B. pseudomallei, which make it less
immunoreactive. 447,99
Although the ability of TLR2 to recognise lipoteichoic acid has led to it
being described as a receptor for the recognition of Gram-positive organisms, 438
TLR2 will also recognise E. coli LPS when associated with lymphocyte
antigen 96 (LY96). 448 Wiersinga reported that TLR2 appears also to recognise
B. pseudomallei LPS in transfected HEK293 cells, 152 but West et al. were unable
49

to replicated this finding using a similar system. 417 There are no clear reasons to
explain the conflicting results, but possibilities include strain differences in LPS
structure 442 and the way in which NF-κB activation was assessed (IL8
production 152 versus a dual-luciferase reporter system 417 ).
Nevertheless, if TLR2 is able to recognise B. pseudomallei LPS, that
interaction is not beneficial to the host, because TLR2 knockout mice have
reduced end-organ damage and improved survival, 152,446 which is contrary to
what would be expected given the results of other mouse models of Gram-
negative infection. 449–452 The reason why Tlr2-knockouts are protected remains
an open question.
1.7.5 Nod-like receptors
The Nod-like receptors (NLR) are cytosolic receptors named for the
nucleotide-binding oligomerization domain (or Nod) that they all possess.
Unlike the TLRs, the NLRs are not membrane-bound: instead, they recognise
antigens present in the cytosol.
The NLRs are divided into two major subfamilies, the NLRCs (NLRC1 to 5)
which possess a caspase recruitment domain and the NLRPs (NLRP1 to 14)
which possess a pyrin domain. 453 A further three genes do not fall into either of
the two major families (NLRA, NLRB1, NLRX1). 453 Like the TLRs, the NLRs are
evolutionarily conserved and are present in species as distantly related as sea
urchins 454 and zebra fish, 455 although they are not present in insects or
worms. 455
The use of animal models to study NLRs is complicated by the fact that
there is not a one-to-one correspondence between human and mouse NLRs:
instead, there are 22 named human NLRs but 34 named mouse Nlrs. 453 For
example, the NLRB family has only one human member (NLRB1) but seven
mouse members (Nlrb1a to f). Similar discrepancies occur in all the families: so,
although there is a human NLRP11, there does not exist a corresponding mouse
Nlrp11. Many of the NLRs remain to be characterised, but are believed to
function as pattern-recognition receptors.
NOD1 and NOD2 (also called NLRC1 and NLRC2) were the first NLRs to
be described 456 and will activate NF-κB directly. 457 Stimulation of RAW264.7
cells (transformed mouse monocytes) with B. pseudomallei causes upregulation
50

of both Nod1 and Nod2. 458 However, no clinical or animal studies of NOD1 or
NOD2 have been reported to date, so their role in melioidosis is as yet
undescribed.
1.7.6 NLRs, caspases, and the inflammasome
The inflammasome is a large macromolecular multiprotein complex first
described in 2002 by Martinon et al. in THP-1 cell lysates (a human monocytic
leukaemia line). 459 The inflammasome was named by analogy to the
apoptosome, the multiprotein complex that drives caspase-9 activation. 460 The
structure of the inflammasome has been confirmed by purifying and assembling
the individual components in vitro. 461
The inflammasome is a complex of three proteins: an NLR receptor, an
adaptor protein, ASC, and caspase-1. Three NLRs are known to initiate
inflammasome assembly: NLRP1, NLRP3 and NLRC4, 460 and all three NLRs
associate with caspase-1 via ASC. 459,462–467
NLRP1 (also known as NALP1 and CARD7) was the first caspase-1
activating receptor identified. NLRP1 recognises muramyl dipeptide (MDP),
which causes a conformational change in NLRP1 that allows it to bind
nucleotides and form oligomers, thus initiating inflammasome assembly. 460
However, the mechanism by which it does so is unknown, because NLRP1 does
not bind MDP. 460 NLRP3 (also known as cryopyrin and NALP3) is stimulated by
LPS, MDP, and bacterial RNA, but assembly of the NALP3 inflammasome
requires a second signal, such as extracellular ATP acting on the P2X7
potassium channel 460 (other second signals include bacterial pore-forming
toxins, silica and asbestos 460 ). NLRP3 will also recognise silica, asbestos and
urate. 460 Neither NLRP1 nor NLRP3 have been studied in the context of
melioidosis.
NLRC4 (also called IPAF and CARD12) will recognise flagellin as well as
the basal body rod component of the B. pseudomallei type 3 secretion system. 468
Salmonella enterica serovar Typhimurium, Legionella pneumophila and
Pseudomonas aeruginosa appear to induce assembly of the NLRC4
inflammasome via the cytosolic delivery of flagellin, which may explain why a
functioning type III or type IV secretion system is necessary for NLRC4
activation. 466,467,469–471
51

Nlrc4-knockout mice have an increased susceptibility to B. pseudomallei
(F. Re, unpublished data 472 ). Unlike NLRP1 and NLRP3 activation, NLRC4
activation is associated with a special type of cell death called pyroptosis, and
pyroptosis is a feature of B. pseudomallei infection of macrophages. 473,474 Taken
together, these data suggest that the NLRC4 inflammasome may be more
important than the NLRP1 or NLRP3 inflammasomes in melioidosis, but this
hypothesis has not yet been addressed experimentally. Pyroptosis is discussed
in more detail below.
Caspase-1 was identified in the late 1980’s and early 1990’s as a cysteine
protease expressed only in monocytes and macrophages, 475–480 and was initially
given the name interleukin 1 converting enzyme or ICE. When the enzyme was
cloned and sequenced in 1992, 480 it was found to be unrelated to any other
known protein, but by 1996, a further nine human and mouse homologues had
been found. 481 In 1996, ICE was renamed caspase-1 482 in recognition of the fact
that it was the first member of this new family to be discovered. There are
currently 12 caspases recognised in humans and mice, 483 but caspase-12 is
inactive in 64% of humans. 484
The caspases were first studied in relation to their role in apoptosis, 481 but
three caspases (human caspase-1, -4, and -5; mouse caspase-1, -11 and -12) 485
have roles in the processing and secretion of pro-inflammatory molecules. The
only caspase for which we have detailed information in melioidosis is
caspase-1. 473,474
Caspase-1 is constitutively present in the cytosol as inactive
pro-caspase-1. 479,480 Activated caspase-1 is responsible for converting
pro-IL1β 480 and pro-IL18 486–489 to their active forms, and will also cleave
pro-caspase-7. 490 It is also responsible for a type of programmed cell death
called pyroptosis, which occurs only in the context of inflammation. 491,492
Breitbach et al. studied experimental B. pseudomallei infection (100 cfu of
intranasal B. pseudomallei) in Casp1-knockout mice on a C57Bl/6
background. 474 In this model, all Casp1-knockouts died ≤72 h after inoculation,
whereas wild-type mice survived ≥6 days. Bacterial loads in lung, liver and
spleen were also higher at 48 h in Casp1-knockouts compared to wild-type
controls. This was confirmed by in vitro experiments, which showed that
intracellular burdens of B. pseudomallei were 10-times higher in Casp1-
52

knockout macrophages. Serum concentrations of IL18 and IFNγ were both
reduced in Casp1-knockout mice.
It has recently been discovered that the most commonly used strain of
Casp1-knockout mouse is also caspase-11 deficient, 492 implying that work done
to date with this mouse may require re-evaluation.
Caspase-3 is the only other caspase that has been studied in melioidosis.
Breitbach reported that caspase-3 was activated only when Casp1 was knocked
out. 474 This is consistent with what we already know: i.e., B. pseudomallei-
infected macrophages die by pyroptosis and not by apoptosis, and the function
of caspase-3 is primarily the induction of apoptosis. 494–496 Nothing further is
known of the role of caspase-3 in melioidosis. Other caspases (caspase-2, -4,
and -8) have also been reported to be upregulated in a BALB/c murine model of
melioidosis, 429 but have not otherwise been studied.
1.7.7 Caspase-1 substrates: IL1β, IL18 and caspase-7
Interleukin 1 was first identified by Gery in 1972 as a soluble factor
produced by mouse peritoneal exudate cells that was capable of stimulating T
lymphocyte proliferation. 497–499 Based on this property, Gery proposed the
name lymphocyte-activating factor or LAF. Dinarello rediscovered the same
cytokine in 1977 and called it leukocytic pyrogen. LAF and leukocytic pyrogen
were shown to be the same in 1979, and renamed interleukin 1. 500 Interleukin 1
was successfully cloned in 1985, which led to the realisation that the IL1
consisted of two distinct but closely related proteins, IL1α and IL1β. 501
IL1β is secreted primarily by monocytes and macrophages and has myriad
functions. 502 Its systemic effects are fever, headache, nausea, myalgia and
fatigue. It has endocrine effects (up-regulation of cortisol and ACTH secretion,
but depression of insulin secretion) and inflammatory effects, including the up-
regulation of chemokines and cytokines (TNFα, IL6, IL8, G-CSF and GM-CSF),
adhesion molecules (e.g., ICAM-1) and the activation of CD4 + T lymphocytes, B
lymphocytes and NK cells. 502
Secretion of IL1β is a two-step process. 460 Pro-IL1β is not present in the
cell in significant quantities, so pro-IL1β must first be transcribed and
translated. Pro-IL1β induction is mediated by NF- κB, and therefore requires
either TLR or NOD2 activation. 465 The second step is inflammasome assembly
53

and cleavage of pro-caspase-1 to its active form, which in turn leads to the
conversion of pro-IL1β to mature IL1β.
In contrast to IL1β, preformed pro-IL18 is present in the cell in significant
quantities, so secretion of IL18 is essentially a one-step process: viz., the
cleavage of pro-IL18 to mature IL18 by caspase-1. 503 This means that the IL18
response always precedes the IL1β responses. Although pro-IL18 is expressed by
a wide variety of cell types, the caspase-1 inflammasome is only assembled in
monocytes or macrophages, 477,486,487 which are therefore the only source of
mature IL18.
Caspase-7 has been called an executioner or effector caspase, 494 because it
cleaves a large set of substrates that ultimately result in the hallmarks of
apopotosis, which are DNA fragmentation and mitochondrial damage. 495,496
Caspase-7 is upregulated in murine melioidosis, 429 but there is currently no
published clinical data on caspase-7 in melioidosis. It seems unlikely that the
primary role of caspase-7 in melioidosis should be apoptosis-related since Sun
et al. have already shown that in the context of B. pseudomallei infection, the
predominant mode of cell death is pyroptosis. In Legionella pneumophila
infection, 504 caspase-7 has been shown to have a role in delivering endocytosed
bacteria to the lysosome. Future studies on caspase-7 using B. pseudomallei
may prove equally illuminating.
1.7.8 Caspase-1, pyroptosis, and cytokine-independent killing
Caspase-1 is responsible for inflammation-related programmed cell death
called pyroptosis, 491,492 which is distinct from apoptosis. Apoptosis is an energy-
dependent process involving the orderly dismantling of a cell by DNA
fragmentation and blebbing, with minimal inflammation. By contrast,
pyroptosis is not energy-dependent: instead, it involves perforation of the cell
membrane, osmotic cell lysis and an accompanying burst of pro-inflammatory
cytokines. 491 Pyroptosis is mediated by a macromolecular complex called the
pyroptosome, which is similar to, but distinct from, the inflammasome. The
pyroptosome is an oligomer, composed of multiple ASC dimers that rapidly
recruit and activate caspase-1. 492,505 A cell only ever assembles a single
pyroptosome.
54

Pyroptosis was originally described in S. Typhimurium infections of mouse
macrophages. 491,492 Miao et al. found that although pyroptosis destroys the
macrophage, this causes the intracellular organisms to spill out into the
extracellular milieu where they are killed. Although Miao’s primary
observations were of S. Typhimurium, he also studied B. thailandensis (which is
closely related to B. pseudomallei) and found that IL1β/IL18 double knockouts
were still capable of controlling infection, but that caspase-1 knockouts were
not. This lead him to conclude that caspase-1 is able to kill intracellular
pathogens by a cytokine-independent mechanism, most likely pyroptosis. 506
Sun et al. showed that B. pseudomallei induces caspase-1-dependent
macrophage death, which is then accompanied by the release of IL1β and
IL18. 473 Although the study predates the first description of the pyroptosome by
two years, Sun found that the process was dependent on the presence of a
functional type 3 secretion system and membrane pore formation, before
proceeding to cell swelling and lysis. 473 These are all characteristic of pyroptosis,
even though Sun was not able to apply that term to his observations. It should
be noted that the precise mechanism by which pyroptosis destroys intracellular
bacteria is unclear: Miao presented data suggesting that the bacteria are
subsequently destroyed by neutrophils, but Breitbach’s in vitro system still
demonstrated an effect despite containing no neutrophils.
The data from cytokine and pattern recognition studies may be
summarised thus: the story begins with B. pseudomallei infecting an immature
monocyte. The infected monocyte is not able to kill the intracellular bacteria
using reactive oxygen species (ROS) or reactive nitrogen species (RNS) because
it has not been stimulated by IFNγ. The monocyte is, however, able to assemble
inflammasomes (and a pyroptosome) under NLRC4 stimulation. The infected
monocyte sends out a cytokine distress signal (TLR-mediated IL12 and NLR-
mediated IL18) before destroying itself and the bacteria by pyroptosis. The
distress signal is picked up by NK and CD8 + cells, which generate and amplify
the critical IFNγ signal. The IFNγ signal then activates intracellular killing
mechanisms in other cells.
55

1.7.9 Monocytes and macrophages
Circulating monocytes may be stimulated to differentiate into any one of
three cells types: dendritic cells, Langerhans cells, or macrophages. 507 Monocyte
differentiation is directed by exposure to different environmental factors
encountered as the monocyte progresses from the peripheral circulation to the
tissues. Exposure to IL4 and granulocyte-macrophage colony stimulating factor
(GM-CSF) causes monocytes to become dendritic cells; 508–510 exposure to IL15
plus GM-CSF causes monocytes to become Langerhans cells; 511 while exposure
to IL6, 512,513 IL10 514 and macrophage colony stimulating factor (M-CSF) 513
causes them to mature into macrophages. Concentrations of M-CSF have never
been measured in acute melioidosis, but plasma levels of IL6 and IL10 are
high, 413,422 which supports a role for macrophages in melioidosis. IFNγ levels are
also high in acute melioidosis, 369,376 and IFNγ is able to direct monocytes to
differentiate into macrophages by stimulating autocrine IL6 and M-CSF
production. 507
In melioidosis, macrophages are the principal orchestrators of the host
cytokine response. They are the generals and neutrophils are the foot soldiers.
Although capable of killing B. pseudomallei, their primary function is not to kill,
but to initiate the host cytokine response (by secreting IL12, IL18 and
TNFα), 360,373,411 in inflammasome activation, 474 and in recruiting neutrophils to
the site of inflammation (by secreting IL1β 515,516* and chemoattractants such as
IL8).
B. pseudomallei is a facultative intracellular pathogen, and monocytes are
an important target. 52,79,80 Macrophages are susceptible to being parasitized by
B. pseudomallei and may be subverted to become incubators for bacterial
replication in human tissues. 52 The effective host response therefore depends on
IFNγ stimulation. Macrophages stimulated by IFNγ are thought to kill
intracellular pathogens by producing a burst of reactive nitrogen intermediates
(RNI) (principally the production of nitric oxide from L-arginine, but also
nitrogen dioxide and nitrous acid) and reactive oxygen species (ROI) (e.g.,
superoxide anion, peroxide, hydroxyl radical and singlet oxygen). 517–520 RNIs
* IL1β is not chemoattractant per se, but it induces the secretion of chemoattractants and the
expression of cell adhesion molecules on the endothelium.
56

and ROIs both appear to have a role in killing B. pseudomallei, 521,522 with the
role of RNIs possible more important than that of ROIs, 521 but they are not the
only mechanism by which stimulated macrophages are able to kill
B. pseudomallei 519,523 and lysosomal killing is probably important. 524
1.7.10 Lymphocytes
Host defences to melioidosis are dependent on macrophage-lymphocyte
interactions, 525 but, unusually, T cell help does not predominate in this
relationship.
In mouse models of melioidosis, the primary source of IFNγ is natural
killer (NK) cells and CD8 + T lymphocytes (at least in the earliest stages of the
host response). 360,373,411 This stands in marked contrast to other pathogens such
as Leishmania major 526,527 and Mycobacterium tuberculosis, 528,529 where IFNγ
is produced primarily by CD4 + T lymphocytes. The lack of a role for helper T
cells in the initial host response to melioidosis means that it lies outside the
Th1/Th2 paradigm within which host responses are usually defined. 530 This may
also explain why clinical epidemiological studies have not found HIV infection
to be a risk factor for melioidosis 180 (unlike other intracellular infections like
leishmaniasis 531,532 and tuberculosis 533,534 ), because HIV depletes CD4 + cells
without depleting NK or CD8 + cells. 535
A role has been found for CD4 + cells in later stages of melioidosis in mouse
models, 411 which leads one to speculate that melioidosis patients co-infected
with HIV may have an increased risk of recurrence, but the small number of
patients co-infected with HIV and melioidosis makes it difficult to collect
enough epidemiological data to examine this question adequately. 180,307,536
1.7.11 Neutrophils
The neutrophil chemoattractant, IL8 (also called CXCL8), is present in
elevated concentrations in clinical melioidosis. 413 Mice do not have IL8, but do
express the related chemokines, CXCL1, CXCL2, and CXCL5, all of which are
elevated in experimental melioidosis. 242 Unsurprisingly, leukocytosis is a
feature of acute clinical melioidosis and this leucocytosis is predominantly a
neutrophilia. 242
The role of neutrophils in experimental B. pseudomallei infection has been
demonstrated convincingly in two mouse studies. 326,537 Bacteria were detectable
57

y flow cytometry ≥24 hours after infection, and the only cell type with which
bacteria were measurably associated were neutrophils, not monocytes or
macrophages. The bacteria were confirmed to be intracellular by confocal
microscopy, but the viability of these bacteria was not confirmed. 537 The ability
of B. pseudomallei to invade and replicate within neutrophils has been
confirmed by multiple independent teams, 79,80,538,539 but the question of
whether neutrophils are a site for B. pseudomallei survival and replication in
vivo remains to be answered.
Neutrophils are an important part of the host defence against melioidosis,
and animals in whom neutrophil function is compromised do worse at
controlling the infection. Easton et al. ablated neutrophils using a monoclonal
antibody against Ly6g (formerly called Gr1) and found that bacterial loads were
much higher after four days than in untreated mice. 326 Secreted
phosphoprotein 1 (SPP1, formerly called osteopontin) is a 33 kDa extracellular
matrix protein that is cleaved by thrombin during inflammation to reveal a
cryptic sequence that is recognised by integrin receptors present on
neutrophils. 540 SPP1 therefore plays a role in the neutrophil recruitment during
melioidosis. In agreement with Easton’s results, van der Windt et al. found that
Spp1-knockout mice had fewer neutrophils at the site of infection, had higher
bacterial loads and worse survival. 242 Plasminogen activator, urokinase receptor
(Plaur, formerly called uPAR or CD87) is expressed by neutrophils and has also
been shown to be necessary for neutrophil recruitment, probably via an effect
on integrin expression. 242 Again, Plaur-knockout mice had a reduced ability to
recruit neutrophils to an area of infection and had a reduced ability to clear
B. pseudomallei, although this time, survival was marginally better in Plaur
knockout mice. 541 This apparent contradiction might be explained by the fact
that Plaur also promotes phagocytosis 541 and B. pseudomallei is capable of
subverting leukocyte endocytosis to further its own survival.
Part of the difficulty with studying neutrophils has been that freshly
isolated neutrophils die within hours of stimulation with live bacteria. It now
appears that this self-destruction is due to a newly described form of
neutrophil-mediated killing, in which neutrophils expel their own chromatin to
form extracellular traps (NET). 542 The role of NETs in the host response to
58

B. pseudomallei remains to be explored, but seem likely to be an important part
of the host response to melioidosis.
1.8 Diabetes and infection
Diabetes mellitus is a heterogenous group of disorders characterized by
chronic elevation of serum glucose levels. The majority of diabetes cases are
primary and idiopathic, but a minority of cases occur secondary to other
conditions. Primary diabetes is divided into types 1 and 2, with type 1 diabetes
being primarily a disease of insulin-deficiency with onset in childhood, resulting
from the loss of insulin-producing beta-islet cells of the pancreas. Type 2
diabetes is associated with obesity, 543,544 increasing age 544–546 and family
history. 547 The primary lesion is not insulin deficiency, but insulin resistance,
with the tissues requiring ever higher concentrations of insulin to achieve the
same effect. Diabetes may also occur secondary to pregnancy, 548 corticosteroid
use, 549–552 Cushing’s disease, 553 pancreatitis, 554 pancreatic resection, 554 cystic
fibrosis, 555 haemochromatosis (both primary 556 and secondary 557 ) and
acromegaly. 558,559
The majority of cases that occur worldwide are due to type 2 diabetes and
there is a problem of under diagnosis, even in developed countries, 560 because
patients often feel well in the early stages of type 2 diabetes and because
screening programmes are not often in place. In 2010, there were an estimated
347 million adults worldwide with diabetes and that number is projected to
increase. 546 Although age standardised rates of diabetes are not increasing in
Southeast Asia, the number of patients with diabetes is increasing due to
population growth and aging. 546 However, the epidemiology of diabetes in
Thailand is complicated by the fact that diabetes in Thailand is associated with
relative insulin deficiency, not just insulin resistance. 775 The reasons for this are
not known.
1.8.1 Diabetes and infection
Patients with diabetes mellitus have an increased risk of developing
infections and sepsis, 561,562 and constitute 20·1–22·7% of all sepsis
patients. 334,563 This association was first observed a thousand years ago by
Avicenna (980–1027 CE), who noted that diabetes was frequently complicated
by tuberculosis. 564 In the pre-insulin era, Joslin noted in a series of 1000 cases
59

that diabetic coma was usually precipitated by infection. 565 Even today,
infection remains an important cause of death in diabetics. 566 Much of the
literature does not distinguish between types of diabetes and regards all
complications as secondary to hyperglycaemia and as independent of diabetes
aetiology.
A small number of conditions are strongly associated with diabetes,
including malignant otitis externa, 567–569 emphysematous pyelonephritis, 570–573
emphysematous cholecystitis, 574,575 Klebsiella liver abscesses, 576 and
rhinocerebral mucormycosis, 577,578 so much so, that when the diagnosis is made
in someone who is not known to have diabetes, the clinician should consider
testing for diabetes.
Most of the aforementioned infections are unusual. Instead, most
infections in patients with diabetes are those that occur also in the general
population. Two population-based studies have proved pivotal to our
understanding of the susceptibilities of patients with diabetes: 561,562 a study of
523,749 Canadians with diabetes and an equal number of matched controls 562
found diabetes increased the risk for cystitis (risk ratio 1·39–1·43), pneumonia
(1·46–1·48), cellulitis (1·81–1·85) and tuberculosis (1·12–1·21). A study of 7,417
Dutch patients with diabetes found a higher incidence of lower respiratory tract
infection (adjusted odds ratios 1·42 for type 1 diabetes and 1·32 for type 2),
urinary tract infection (1·96 and 1·24), and skin and mucous membrane
infection (1·59 and 1·33). 561 The association between diabetes and tuberculosis
was re-confirmed by a recent meta-analysis of 16 studies spanning a 50-year
period. 579
Although diabetes mellitus is implicated in susceptibility to infection, its
influence on the subsequent clinical course and outcome is less clear. Some
studies have shown an association with increased mortality, 580–583 others found
no effect, 563,584–592 while still others found improved survival. 574,575,593 The
largest of these (12·5 million sepsis cases) 574 found that patients with diabetes
were less likely to develop acute respiratory failure and linked this to two
previous studies that found diabetics seem protected from acute lung
injury. 594,595 The largest single study to show an adverse effect of diabetes on
mortality in sepsis was conducted in 29,900 Danish patients with community-
60

acquired pneumonia and found that patients with diabetes had a higher risk of
mortality (OR 1·2). 582
The reasons for different outcomes between these studies are unclear, but
may relate to differences in study population, varying outcome measures,
differences in statistical analysis and in diabetes drug prescription habits
between countries. 329 Population-based studies are less prone to selection bias
compared to hospital-based studies, but more detailed clinical information is
usually available in hospital-based studies. In terms of outcome measures,
studies with outcomes at longer time points (e.g., 6 months versus 28-day
mortality) are more likely to find informative differences, but are much more
difficult to conduct. 596 Observational studies often make use of multivariable
regression techniques to correct for confounders (a common, but incorrect,
approach to model-building is to include all measured parameters and then
remove parameters on the basis of p-value). Overadjustment or unnecessary
adjustment for variables that are not confounders can produce biased or
spurious results (these issues are further discussed in Chapter 2). 597 Patients
with diabetes also have multiple co-morbidities that may worsen outcomes: it is
debatable whether these co-morbidities should be adjusted for, since many are
caused by diabetes itself and therefore, by definition, are not confounders. 598
Nevertheless, a number of studies have adjusted their results for these
comorbidities, regardless of the validity of these adjustments. 581,582,584
1.8.2 Diabetes and the host response
In 1904, Lassar suggested that high levels of glucose may drive infection by
serving as a nutrient source for bacteria, 599 but in 1911 Handmann showed that
glucose supplementation did not enhance bacterial growth, 600 and proposed
instead a defect in immune function, which Da Costa and Beardsley
demonstrated in 1907. 601 The subsequent literature on this topic is complicated
by the fact that different techniques have been used over the years, and gaps of a
decade or more may separate experiments, making it difficult to compare
results. Most studies have shown defects in neutrophil function, with good
evidence for abnormalities in adhesion, chemotaxis and intracellular killing, but
evidence for a phagocytosis defect are contradictory. The evidence that
neutrophil defects are solely responsible for the increased susceptibility of
61

diabetics to infection is equivocal. 602 There is good evidence that humoral
responses in diabetics are poorer, and may play a larger role than previously
recognized.
1.8.3 Diabetes and general markers of inflammation
Diabetes is associated with elevations in CRP, 603 TNFα, 604 IL6 603 and
IL8, 605 but no differences are seen in circulating cell surface markers or
coagulation markers between patients with and without diabetes in the context
of sepsis.
In a cohort of 1,799 patients with community-acquired pneumonia
(CAP), 606 concentrations of pro-inflammatory cytokines (TNFα, IL6 and IL10),
coagulation (antithrombin, Factor IX and thrombin-antithrombin complexes)
and fibrinolysis (PAI-1 and D-dimer) biomarkers were similar in subjects with
and without diabetes at presentation and in the first week of hospitalization. 606
In addition, monocyte expression of CD120a, CD120b, HLA-DR, TLR4 and
TLR2 on monocytes was not different between the groups. 606 These results are
consistent with a cohort study of 830 sepsis patients, in whom plasma
concentrations of IL6 and TNFα were elevated to the same extent in patients
with and without diabetes, both at admission and at follow-up. 563 In this second
study, diabetes was not found to exacerbate the known procoagulant response
seen in sepsis. 563 Since sepsis and diabetes both induce a proinflammatory and
procoagulant state and since both interfere with the host response, the lack of a
strong influence of diabetes on the proinflammatory and coagulation pathways
during sepsis is remarkable. Preclinical studies in healthy volunteers have
shown that acute hyperglycæmia and insulin resistance may both directly
influence inflammation and coagulation, 607,608 but these changes may not be
detectable on the background of the much larger abnormalities attributable to
sepsis. There is also evidence that local responses may be impaired in diabetes:
e.g., levels of urinary IL6 and IL8 are lower in diabetic women with
bacteriuria. 609 Endothelial activation has been implicated in the pathogenesis of
sepsis 610 and diabetes is itself known to activate endothelium. A recent study of
207 sepsis patients (of whom 30% had diabetes) showed that markers of
endothelial cell activation (plasma E-selectin and sFLT-1) were higher in
diabetes. 611
62

1.8.4 Diabetes and neutrophil adhesion
Recruitment of neutrophils to a site of inflammation requires endothelial
adhesion followed by transmigration and exit from the circulation, a process
requiring the expression by neutrophils of integrins (e.g., CD11a/CD18 and
CD11b/CD18), 612,613 which then bind to endothelial cell adhesion molecules
(e.g., intercellular adhesion molecule 1, or ICAM-1 614–616 ).
A study in which neutrophils were harvested from 26 patients with
diabetes and an equal number of controls demonstrated that adhesion to bovine
aortic endothelium was increased for neutrophils from diabetics, but only if the
endothelium was also incubated with plasma from patients with diabetes. 617
Increased adhesion appears to be due to both an increase in expression of
integrins by diabetic neutrophils and of adhesion molecules by endothelium.
Diabetic neutrophils have increased expression of CD11b and CD11c, 618 and
glucose itself appears able to stimulate the expression of ICAM-1 by endothelial
cells, 614–616,619–621 possibly via an osmotic effect. 620,621
1.8.5 Diabetes and neutrophil chemotaxis
Chemotaxis is the ability of neutrophils to detect and move towards a
chemical stimulus. Studies of chemotaxis may be divided by technique: those
using the two-chamber Boyden technique 622 have produced conflicting
results, 623,624 but the subagarose technique 625 (which includes a negative
control, which Boyden’s technique lacks) have reproducibly shown a defect in
diabetes. 618,626
1.8.6 Diabetes and neutrophil phagocytosis
Phagocytosis is the engulfment and ingestion of foreign bodies by a cell,
allowing neutrophils to remove and destroy pathogens.
The evidence for a defect in phagocytosis in diabetes is contradictory, with
some reporting a defect, 627–630 but others not. 618,631 These inconsistencies may
be attributed to differences in methodology: neutrophils will not phagocytose
unopsonised particles, so bacteria and cells need first to be incubated with
serum containing C3b or IgG. Many studies have used autologous serum, 627–630
but those that have used a standard serum or opsonin have found no
defect. 618,631 In 1976, Bagdade found that phagocytosis of Streptococcus
pneumoniae was reduced in neutrophils recovered from eight patients with
63

poorly-controlled diabetes, but this defect improved with diabetes treatment. 629
Notably, control neutrophils incubated with serum taken from patients with
diabetes also demonstrated a defect in phagocytosis, implying that the defect is
in fact due to defective opsonisation and not to a deficit in neutrophil
phagocytosis per se: in other words, the defect is humoral. In 1984, Davidson et
al. studied ingestion of Candida guilliermondii by neutrophils from 11 patients
with diabetes and found that phagocytosis was reduced. However, if pre-
opsonized yeast cells were used, then phagocytosis was no different from
controls, again suggesting that a humoral defect must exist. 631 Delamaire et al.
used a single control serum for all samples to remove the possibility of a
difference in opsonisation, 618 convincingly demonstrating that no phagocytosis
defect exists in diabetic neutrophils.
1.8.7 Diabetes and neutrophil-mediated killing
Neutrophils have two distinct mechanisms for killing bacteria,
intracellular and extracellular. Phagocytozed bacteria are killed by superoxide
anions and other oxygen-derived species. Culture-based methods have
demonstrated a defect in intracellular killing of Staphylococcus aureus, 626,632,633
Streptococcus pneumoniæ, 628,634 and Candida albicans. 635 More recent studies
have confirmed this finding used chemiluminescence methods, 636–639 a superior
method compared to culture, because it separates the effect of phagocytosis
from that of intracellular killing. The killing defect cannot be corrected by
incubation with normal serum, 633 suggesting that it is cellular in origin. The
defect improves with glycaemic control. 638
Neutrophils are also able to kill bacteria extracellularly by expelling
chromatin, which combines with granule proteins to form NETs. 542
Interestingly, β-hydroxybutyrate (a ketone body present in diabetic
ketoacidosis) has been shown to inhibit the formation of NETs, 640 but the
relevance of this finding to patients remains to be demonstrated.
1.8.8 Diabetes and monocyte function
Monocytes in diabetes have been less well studied than neutrophils, but
also appear to have defects of chemotaxis 641 and phagocytosis. 642,643 Adhesion to
endothelium is also enhanced. 644,645 In contrast to neutrophils, intracellular
killing seems to be enhanced. 646 Monocytes obtained from 24 diabetic patients
64

produced similar amounts of TNF-α when compared to healthy controls when
stimulated with lipopolysaccharide (LPS), but IL-6 levels were higher in
patients with type 1 diabetes. 647
1.8.9 Diabetes and lymphocyte function
Few studies have looked at the effect of diabetes on lymphocyte function.
One measure of lymphocyte function is transformation in response to a mitogen
or bacterial antigen. Studies containing acidotic patients appear to find that
responses are diminished 648,649 and that correction of the acidosis leads to
prompt resolution of the defect, 649 but more recent studies have found deficient
proliferative T-cell responses even in treated patients. 639,650 Diabetic T-cells
express higher levels of CD152, a down regulator of the immune response. 651
Three other studies failed to find a defect in lymphocyte function. 624,652,653
1.8.10 Humoral defects
In 1907, Da Costa and Beardsley found that sera from diabetes patients
were less able to opsonize S. aureus compared to sera from controls. 601,654 In
1973, Farid and Anderson surveyed 46 patients and found that IgG levels were
lower in insulin-treated diabetics, but not in patients on oral treatments or diet
alone. More recently, a study of 66 patients with type 1 diabetes demonstrated
that total IgG levels were lower in patients with uncontrolled diabetes as
measured by HbA1c. 655 Also, the apparent defect in neutrophil phagocytosis
appear to be humoral and not cellular in origin (see above).
The best evidence for a humoral defect in diabetic patients comes from
vaccine studies. It was described as early as 1930 that deficient agglutinin
responses are seen in diabetic patients after subcutaneous typhoid
vaccination. 656,657 Multiple studies have shown that patients with diabetes are
less likely to mount a protective antibody response to hepatitis B
vaccination, 658–661 leading some authorities to recommend routinely adding a
booster dose to the standard regimen for patients with diabetes. 658,662 The
literature on influenza vaccination is more mixed (reviewed by Brydak and
Machala 663 ). Pozzilli et al. looked at 52 diabetic patients and found fewer
activated lymphocytes in patients with type 2 diabetes following influenza
vaccination, but no differences in antibody responses. 664 Muszkat et al.,
65

studying a more elderly population, found lower antibody responses in patients
with type 2 diabetes. 665
Diabetes is also associated with a waning in the duration of protection
afforded by tetanus vaccination although the initial response appears to be
normal. 666,667 Patients with diabetes appear to respond well to pneumococcal
polysaccharide vaccine, 668 although there are no studies studying duration of
protection in diabetic patients. There are also no studies specifically linking
humoral responses in sepsis to diabetes.
1.8.11 Complement abnormalities
Inherited deficiencies of component 4 (C4) have been implicated in the
pathogenesis of type 1 diabetes, 669–671 but whether this contributes to
susceptibility to infection in type 1 diabetics is not known. By contrast, obesity
and elevated insulin levels (as occur in type 2 diabetes) appear to be associated
with elevations in C3. 672 Karlsson et al., looking for biomarkers for maturity-
onset diabetes of the young (MODY), found that complement C5 and C8 are
both elevated in diabetes regardless of etiology, 673 a possible mechanism for
these abnormalities being that complement activation can be driven by glycated
immunoglobulins. 674 One explanation for why diabetic sera are less able to
opsonize bacteria may be that glucose attacks the thiolester bond of
complement C3, thus preventing it from binding to the bacterial surface. 675
1.9 Diabetes and melioidosis
1.9.1 B. pseudomallei and the diabetic neutrophil
Diabetes is the most important risk factor for melioidosis, and the
epidemiological evidence for this has been reviewed above. We also know that
the host response to melioidosis is critically dependent on neutrophils, 326 and
diabetes is known to cause defects in neutrophil function, as reviewed above.
The effect of diabetes on the neutrophil response to B. pseudomallei
infection has been explored by Chanchamroen et al. 539 Neutrophils were
collected from 56 patients with diabetes who were otherwise healthy and
compared to 36 healthy controls. Cells were then infected with B. pseudomallei
K96243. All study subjects were from Northeast Thailand.
66

Chanchamroen demonstrated that B. pseudomallei is able to persist
intracellularly in neutrophils. She also reported impairment of phagocytosis,
chemotaxis and an increased rate of neutrophil apoptosis. Her results are
difficult to interpret due to a number of limitations in the study design as
follows:–
Chemotaxis was studied using the two-chamber method, 622 which is
inferior to the subagarose technique 625 because it has no negative control and
cannot distinguish between specific and non-specific migration. Despite this,
she was still able to show a chemotactic defect in diabetic neutrophils.
Phagocytosis was measured using autologous serum, a method which we
now know to be severely confounded by humoral effects. The study appeared to
show increased intracellular killing by diabetic neutrophils, which is a totally
unexpected finding. The intracellular killing assay is severely confounded by the
fact that phagocytosis was poorer in diabetic neutrophils, which means the
initial number of intracellular bacteria was smaller in the diabetic neutrophils.
This was not corrected for in the analysis, which in turns means that no
meaningful comparison between diabetic and non-diabetic neutrophils was
possible. The same study demonstrated that B. pseudomallei appeared to delay
apoptosis in healthy neutrophils but not in diabetic neutrophils: however,
interpretation of this result is also problematic due to the same confounding
factors.
1.9.2 B. pseudomallei and the diabetic macrophage
Williams et al. studied peritoneal macrophages and bone marrow-derived
dendritic cells harvested from C57Bl/6 mice made hyperglycaemic by treatment
with low-dose streptozocin (STZ). 676 In this model, mice were treated with
55 mg/kg STZ intraperitoneally for six days; control mice were treated with
buffer only. Mice were considered diabetic if their blood glucose levels were
≥13 mM for 9 days. Mice in the ‘acute diabetes’ group were sacrificed 19 days
after the first STZ injection, while mice in the ‘chronic diabetes’ group were
sacrificed 70 days after the first infection.
Dentritic cells were harvested by culturing bone marrow for 10 days with
GM-CSF. Peritoneal macrophages were harvested after injection of the
inflammatory agent, Brewer’s thioglycollate. Phagocytosis was measured using
67

the ability to internalise fluorescent beads. Intracellular killing was measured by
lysing the cells with 0·1% Triton-X, then culturing the lysate on Ashdown’s agar.
Colony counts at 18 hours were expressed as ratios of the number of bacteria
internalised at 4 hours, thus adjusting for differences in phagocytosis. Cytokine
expression was measured by quantitative PCR.
Williams found no differences in cells collected from the acute diabetes
group when compared to the controls, but in cells collected from the chronic
diabetes group, intracellular survival of B. pseudomallei was greater than in
controls. She also found that expression of IL11β, IL12, IL18 and TNFα were
depressed in the chronic diabetes group compared to the control group.
1.9.3 Melioidosis in murine models of diabetes
Hodgson has published a murine model of melioidosis in type 2 diabetes,
using the BKS.Cg-Dock7 m +/+ Lepr db /J mouse (also called the db/db mouse). 677
This inbred strain was originally created in 1966 from a C57Bl/KsJ
background. 678 Homozygotes for the spontaneous leptin receptor mutation,
Lepr db , become hyperinsulinaemic at 10–14 days of age, obese at 3–4 weeks,
hyperglycaemic at 4–8 weeks and die at around 10 months. 679 Non-diabetic
litter mates heterozygous for the Lepr db mutation were used as controls.
Diabetic mice were more susceptible to B. pseudomallei infection than age
and sex-matched controls: following a subcutaneous challenge of 4·6 × 10 6 cfu,
all mice in the diabetic group were dead by day 4, whereas the control mice all
survived until day 11. In this study, no differences in neutrophil oxidative burst
were found between diabetics and controls, but no other test of neutrophil
function was performed. This is in contrast to another study by a different group
using S. aureus that showed impaired oxidative burst in db/db neutrophils. 680
Hodgson found no differences in bacterial loads in liver or spleen, but did
find higher loads in subcutaneous tissue. Concentrations of TNFα and IL1β
mRNA in the spleens of db/db mice were higher than controls. In vitro studies
of peritoneal macrophages harvested from db/db mice showed that
B. pseudomallei was better able to survive in db/db macrophages and that
db/db macrophages produced lower levels of nitric oxide. Hodgson therefore
attributed the increased susceptibility of diabetic mice to a defect in
macrophage responses.
68

1.9.4 Melioidosis and diabetes in the clinic
It has been recognised that once melioidosis has been acquired, patients
with diabetes paradoxically seem to do better. In a prospective observational
study of 755 melioidosis patients, diabetes was independently associated with
survival (adjusted odds ratio for mortality 0·36, 95% confidence interval 0·25–
0·50). 681 This is at odds with what we know about the effect of diabetes on the
host response, which would lead us to predict that outcomes should be poorer in
diabetes. This finding is not unique to melioidosis, and the conflicting data in
the literature on the influence of diabetes on sepsis outcomes has been
discussed above.
1.10 Outline of the work
In this dissertation, the link between diabetes and melioidosis is explored
using classical epidemiological techniques (Chapter 2), and it is shown that this
link is attributable to the use of glibenclamide prior to hospital admission for
melioidosis and not to diabetes per se. The host response to melioidosis is
explored by studying gene expression in peripheral blood leukocytes
(Chapter 3), and glibenclamide is shown to have an anti-inflammatory effect in
patients with melioidosis. The effect of glibenclamide is then described in a
mouse model of melioidosis (Chapter 4), in which it is shown that glibenclamide
reduces interleukin-1β secretion.
69

2 Glibenclamide, not diabetes, protects
from mortality in melioidosis
2.1 Introduction
2.1.1 The melioidosis cohort at Sappasithiprasong Hospital
The Wellcome Trust was founded in 1936 under the will of Sir Henry
Wellcome, co-founder of the pharmaceutical firm, Borroughs-Wellcome, who
established tropical medicine research laboratories in London and Khartoum
following visits to Ecuador and The Sudan. The Wellcome Trust Unit in
Bangkok was started in 1979 as a long term collaboration between Oxford
University and Mahidol University by Professor Sir David Weatherall, Professor
David Warrell, Professor Emeritus Chamlong Harinasuta and Professor
Emeritus Khunying Tranakchit Harinasuta with support from the then director
of the Wellcome Trust, Dr Peter Williams.
The Unit’s involvement in melioidosis research began when the then
director of the Wellcome Trust Unit, Professor Nicholas White, was approached
by a physician from Ubon Ratchathani, Professor Wipada Chaowagul, who had
noticed increasing numbers of melioidosis cases in her practice. The laboratory
at Sappasithiprasong Hospital (Figure 11) was established by David Dance and
Nicholas White in 1986.
Ubon Ratchathani is a province in Northeast Thailand (Figure 12),
629 km to the east of Bangkok. It is bordered by Laos to the east (from which it
is separated by the Mekong River) and Cambodia to the South. Ubon
Ratchathani was founded in 1786 as a vassal state of Siam and became a
province of Thailand in 1882. It was the largest province in Thailand until
Yasothon (1972) and Amnat Charoen (1993) gained status as independent
provinces.
70

Figure 11. Sappasithiprasong Hospital, Ubon Ratchathani.
Note.—The front façade of Sappasithiprasong Hospital, Ubon Ratchathani.
Sappasithiprasong is a 1000-bedded tertiary referral centre for the province of Ubon
Ratchathani. ©2008 GCKW Koh.
Ubon Ratchathani province has an area of 16,112 square kilometres and a
population of 1·8 million (2009 data). It is subdivided into 25 districts (called
amphoe) and is further subdivided into 219 subdistricts (thambon). The largest
city is Ubon Ratchathani city (population 85,000), followed by Warin Chamrap
(population 31,000). The main industry is rice farming. The climate is tropical,
with a rainy season that runs roughly from June to October, and a dry season
that runs roughly from November to May.
Sappasithiprasong Hospital is a 1000-bedded provincial hospital. It is the
largest hospital in the region and a tertiary referral centre for 25 community
hospitals. Data from all patients who were culture-positive for B. pseudomallei
have been collected prospectively from the establishment of the Wellcome Trust
laboratory at Sappasithiprasong until 2007. The data from this cohort has been
an important source of information on the epidemiology of melioidosis. 125,169,170
71

Figure 12. Ubon Ratchathani province.
Note.—Ubon Ratchathai province is marked in red. The province is bordered by the
Mekong River to the east, which separates it from Laos, and Cambodia to the south. Its
internal boundaries are formed by the provinces of Yasothon and Amnat Charoen to the
north and Srisaket to the west. ©2009 NordNordWest, Wikimedia Commons.
It has previously been observed in Ubon Ratchathani, that although
patients with diabetes are more likely to develop melioidosis, they are also less
likely to die from it. 681 This finding was made incidentally in a study of the
prognostic value of semi-quantitative urine cultures in melioidosis and diabetes
was not the primary exposure of interest. We sought to confirm this association
between diabetes and survival in a separate cohort of patients from the same
hospital using causal reasoning to identify possible confounders, and then to
adjust for these confounders using multivariable logistic regression models. We
hoped our methods would be rigorous enough either to confirm convincingly
the effect of diabetes, or else to identify alternative explanations for the
observed effect.
72

2.1.2 Logistic regression
Logistic regression has two common uses. The first is to estimate the
probability of a particular outcome given some set of risk factors. An example of
this first use is the APACHE score, which combines multiple clinical and
laboratory parameters to provide an estimate for ICU mortality. A second use is
to adjust for confounding in an estimate of the effect of an exposure on
outcome. It is for this second use that we employed logistic regression in the
cohort study.
Figure 13. The logistic function.
Note.—The logistic function can take any value as its input, but its output is
constrained to lie between 0 and 1. More formally, the domain of the logistic function is
(–∞, +∞), and its co-domain is (0, 1). ©2008 Qef, WikiMedia Commons.
The logistic equation may be defined as,
! = !
!!! !!,
where
73
Equation 1
! = ! ! + ! !! ! + ! !! ! + ⋯ + ! !! !. Equation 2
The main value of the logistic function is that it may take as its input, !,
any value for ! from −∞ to +∞, but its output, !, is constrained to lie between 0
and 1 (Figure 13). This is why the function is given the name, ‘logistic’: it is
useful for circumstances where the outcome may take only two values, e.g.,
1/0, yes/no, true/false. The value of ! may then be interpreted as the probability
of the outcome being ‘yes’ or ‘true’, given a particular value for !.

The logistic function may take any number of inputs, !, (in the equation
above, these inputs are numbered, ! !, ! !, …, ! !), however, there can be only one
output, ! . The variables, ! and ! , may be called by a variety of names,
depending on the context (Table 7).
x y
input output
abscissa ordinate
independent variable dependent variable
co-variate outcome
exposure disease
risk factor
Table 7. Alternative names for x and y in multivariable regression
models.
Note.—The variables ! and ! can take a great many names depending on the context.
The terms input/output and abscissa/ordinate are used mainly be mathematicians.
Statisticians tend to speak of independent/dependent variables, while epidemiologists talk
of exposure/outcome or risk factor/disease. The term ‘disease’ is very inexact, because
depending on the context, different diseases may appear as x or y variables.
The logistic equation has been presented here in a form that emphasises
its link to linear regression. Equation 2 is in fact the linear regression model. In
Nelder and Wedderburn’s method of generalised linear models, 682 logistic
regression is related to linear regression via equation 1, which is the link
function.
The logistic function, in its simplest form, was first described by a Belgian
mathematician, Pierre-François Verhulst, in 1838, 683 as the exact solution to the
first order differential equation, † ! !"
!"
= !!"(1 − !).
† !
!"
Here is a simple demonstration that ! =
!!! !! is a solution for = !(1 − !).
!"
Note that ! may be any constant, so for the purposes of this analysis we may set it to 1 without
affecting the validity of the results.
74

The paper of 1838 was brief, but a fuller account was given in 1845, 684 in
which, for the first time, Verhulst calls the solution, la courbe logistique or ‘the
logistic curve’. Verhulst applied the equation to model population growth, in the
situation where that growth is initially exponential, but then encounters
resistance (that is, the exhaustion of resources). Berkson was the first to
advocate its use in bioassays 685 (a variation of this curve is commonly used to
analyse ELISA results) and other catalytic processes. 686 Berkson also introduced
the term ‘logit’ (by analogy to ‘probit’). 686 The logistic function is therefore also a
good model for the market penetration of iPhones, which are also a catalytic
process: growth is initially exponential, but then slows as the market becomes
saturated. 686
The use of the logistic function in case-control studies arose naturally. If ! !
is defined as the odds, then ! is the log odds, and the linking function
(Equation 1) is then mathematically identical to the risk. The application of the
logistic function to binary data may be attributed to Cox, who published a series
of papers in the 1960’s, culminating in an influential textbook on the subject in
1969. 687 Two events allowed the expansion of logistic regression in
epidemiological research during the 1990’s: the first was the availability of
computer software packages that could fit the models automatically (fitting
curves by hand is laborious and takes weeks of tedious calculation). The second
was the publication of a textbook by Hosmer and Lemeshow in 1989, 688 which
provided guidance on building models while omitting mathematical details that
could be left to the computer.
Mathematically, the minimum number of precisely specified data points
needed to fit a logit model is one plus the number of inputs to be estimated. In
Verhulst’s original 1838 paper, he fit curves for the populations France, England
and Russia based on three points only each. The naturally occurring variation in
clinical data means this is never possible in practice. Instead, the number of
inputs that may be included in a model is limited by the amount of data
available to fit the model.
There is no consensus on how much data is ‘enough’. Computer
simulations suggest that for ! variables, a minimum of 10n data points are
Starting with ! = ! !!
!!! !! =
!!! ! , multiplying out the fractions yields ! + !!! = ! ! .
Differentiating, gives us , !"
!"
!"
+ !!
!" + !!! = ! ! . Re-arranging, !"
!"
75
= !!
!!! !
1 − ! = !(1 − !).

needed for each input variable. 689 Some authorities have suggested that is too
stringent; 690 but some have suggested instead much larger numbers, of the
order of 4 n . 691 If the figure of 10n is right, then a model with ten variables only
needs a study of 50 patients; using a figure of 4 n , the same study needs 1024
patients. These uncertainties mean that sample size calculations for
multivariable models are often done simply for the univariable analysis relating
exposure and outcome, and not for the multivariable analysis.
2.1.3 Adjusting for confounding
Logistic regression provides a powerful method for adjusting for
confounding. In a cohort study, where the natural metric for risk is the risk
ratio, odds ratios are still frequently reported, because these can be adjusted for
confounders using logistic regression.
The classical method of adjusting for confounding is the Mantel-Haenszel
method. Briefly, one stratifies the data by each confounder, calculates the
estimate for each stratum, then combines the stratified estimates into a single
summary estimate.
The method is best illustrated using a hypothetical example. Consider the
effect of diabetes on mortality (Figure 14). A worked example of stratification
using invented data is presented in Table 8.
The Mantel-Haenszel method has two major problems: the first is that it is
not good at handling adjustment for several confounders. The second is that the
method will only allow categorical confounders, so continuous data first needs
to be transformed into categorical data before it can be used, which in turns
leads to inefficiencies (in that one has to discard information before performing
the analysis). Using a logistic regression model to adjust for confounding
addresses many of these concerns. The model easily accepts continuous
variables as inputs and is more forgiving of sparse data.
76

Figure 14. The influence of diabetes on mortality is confounded by
gender.
Gender Diabetes
Mortality
Flow of time
Note.—The effect of diabetes on mortality is confounded by the effect of gender, because
gender is associated with both diabetes and with mortality and does not lie on the causal
pathway between diabetes and mortality.
The thick arrows in the graph indicate the flow of causation and must always follow the
direction of time, because effect always follows cause. Confounding occurs when one is
able to follow a path from exposure to outcome that goes against the direction of
causation at some point (i.e., against the flow of time). This is known as a backdoor path.
In this example, the effect of interest is the effect of diabetes on mortality (black arrow);
there is however, a backdoor path linking diabetes and mortality that runs through gender
(blue arrows). Gender is therefore a confounder for the effect of diabetes on mortality.
One cannot infer causation from data alone, in part, because the data contain no
information about time sequence. Information about time ordering must be added to the
analysis by the investigator. In epidemiological studies, the person with this information
is the epidemiologist, not the statistician. It is therefore never the statistician’s
responsibility to identify confounding.
77

A. Fictional population
Male Female
Diabetic Not diabetic Diabetic Not diabetic
Died 450 10 200 360
Total 900,000 100,000 100,000 900,000
B1. Matched case control study: unstratified
Diabetic Not diabetic
Died 650 370
Alive 470 550
B2. Matched case-control study: stratified by gender
Male Female
Diabetic Not diabetic Diabetic Not diabetic
Died 450 10 200 360
Alive 414 46 56 504
Table 8. A hypothetical study population.
Note.—Table A presents data from a purely hypothetical population categorised by three
parameters: gender (male/female), diabetic (yes/no) and mortality (died/survived). Table B
shows data from a case-control study drawn from that hypothetical population. The study
investigators in our idealised study were able to recruit every patient who died, and the
controls are perfectly frequency matched by gender.
A: the odds ratio for a person dying if he or she has diabetes is 5.0
= (450×100,000)/(10×900,000) = (200×900,000)/(360×100,000).
B: In B1, the odds ratio as estimated from the unstratified data is (650×550)/(370×470) =
2·1. This value is clearly incorrect, because in this hypothetical example, we know that
the true population odds ratio is 5·0, because that is what we set it to be. In this case, we
know that the estimate is confounded by gender. In B2, the same data has been divided
into two strata, Male or Female. This removes the effect of confounding, because every
person in each stratum has the same gender (either Male or Female). In the Male stratum,
the odds ratio is (450×46)/(414×10) = 5·0; In the Female stratum, the odds ratio is
(200×504)/(56×360) = 5·0. Combining the odds ratios in this simple example yields a
summary odds ratio of 5·0, which is the correct value.
78

Aside from demonstrating how to adjust for confounding, this example demonstrates that
matching on a confounder in a case-control study is not sufficient to remove the
confounding. The final estimates from a matched case-control study must still be adjusted
for all confounders that were matched, and failure to do so will produce erroneous
results. It is even possible to demonstrate that matching for a variable that is not a
confounder in a case-control study is capable of introducing confounding into the results.
Many authors have therefore argued against matching in case-control studies, and many
case-control study designs are now unmatched. This example is adapted from Rothman
1986. 692
To adjust for more confounders, the data must be divided into strata for
every possible combination of variables. For example, to adjust for the effect of
obesity as well as a the effect of gender, a minimum of four strata would be
needed: obese/male, obese/female, non-obese/male, non-obese/female. The
number of strata required therefore increases exponentially with the number of
confounders. If there are too many confounders to adjust for, there will not be
enough data to fill every stratum and some strata will be empty or partly empty.
The calculation becomes impossible for empty strata, because the estimated
odds ratios will be either zero or undefined (division by zero).
The statistical package, Stata (StataCorp, College Station, Texas), fits
logistic regression models using maximum likelihood estimation, which is
better able to cope with sparse data. The command logistic reports odds
ratios, while logit reports log odds ratios: the models produced by both
commands are identical.
Logistic regression can be extended to accommodate outputs with more
than two levels (multinomial or multivariate logit modelling), or situations
where the inputs are not independent (conditional random field modelling).
These are implemented in Stata as blogit and glogit, respectively.
2.1.4 Identification of confounders and causal reasoning
Informally, a confounder is one that accounts for all or part of an apparent
association between the exposure and the outcome and is not itself affected by
the exposure. One definition of a confounder is that a confounder is any
parameter associated with both exposure and mortality, but which does not lie
79

on the causal pathway between exposure and mortality. 598 The identification of
confounders is the business of the epidemiologist and not the statistician,
because it requires knowledge about the data that cannot be inferred from the
data itself. 693 There is no statistical test for confounding, so any method based
solely on p-values is incorrect. 694
The greatest challenge facing the model-builder is the correct
identification of confounders to include in the model. There exists a popular
misconception that adjusting for more variables is a good thing. Few data sets
are large enough to allow adjustment for every possible confounder and
unnecessary adjustment wastes power: in other words, a true effect can be made
to spuriously disappear by adjusting unnecessarily for something that is not a
confounder. 597,688,695
There is an even more serious problem with unnecessary adjustment.
Adjusting for a parameter which is not a confounder may introduce
confounding and hence produce biased results. 597,693,696
When logistic regression models were first introduced into clinical
research in the 1970’s and 80’s, model fitting had to be performed by hand.
Starting in the early 1990’s, computer software packages became available that
would fit logistic regression models in milliseconds, which has led to an over-
dependence on computer-led methods of model building, principally, the step-
wise selection of variables.
Generally speaking, step-up procedures add variables one by one on the
basis of p-value (smallest p-value first), while step-down procedures throw all
the variables into the model and then remove them one by one according to p-
value (largest p-value first). 697 The attraction of these data-driven approaches is
that they remove the necessity for careful thought and will invariably produce
models with excellent fit. Unfortunately, it has been repeatedly shown that
models produced in this way give overoptimistic estimates for p-values and
seldom give reproducible results 688,695 (a problem known as ‘over-fitting’).
A more serious problem with stepwise methods is their validity. The p-
value reported (whether calculated by the Wald test or by the likelihood ratio
test) measures the value of that risk factor in predicting the outcome. Stepwise
methods are therefore only valid (if they are ever valid) for the purpose of
building a predictive model. When models are built to adjust for confounding,
80

the p-values are irrelevant, because the hypothesis they test is unrelated to
whether the variable is or is not a confounder. 698
The selection of variables for inclusion in a model is always dependent on
external information, 693,699 and is only weakly dependent on information
contained in the data itself. An important example of this external information
is a knowledge of the temporal relationship between variables. In fact, this is
probably the single most important piece of information available to the
investigator, for the single reason that effect must always follow cause. If
variable A follows variable B in time, then variable A can never be a confounder
for variable B, because confounders must always precede in time, or at least be
roughly contemporaneous with the variables they confound (thus allowing for
the possibility that variable A might precede variable B).
The simplest tool for identifying confounders is the conceptual
hierarchy 700 (Figure 15). In the hierarchical conceptual framework that we built
for this cohort study, time flows from the top of the chart to the bottom. In our
cohort study, the exposure of interest was diabetes (Figure 15, level 2) and the
outcome of interest was death (Figure 15, level 6). Any one factor can be
confounded only by other factors in the same level or in the levels above.
Factors in lower levels must be either in the causal pathway or be common
effects: they cannot be confounders.
The question then arises: What is the effect of adjusting for variables on
the causal pathway? All diabetes drugs must lie on the causal pathway between
diabetes and death, the reason being that they are only prescribed after the a
diagnosis of diabetes is made and are only prescribed in the context of diabetes.
However, the choice of drug may be manipulated by the clinician. Adjusting for
drug therapy allows us to ask the question, “what if we did not prescribe this
drug?” It provides an elaboration of what is going on and it permits us to look
for possible mechanisms. It is important to note that while this operation is
mathematically identical to adjustment for confounding, the interpretation of
the result is completely different from adjusting for a confounder.
81

1 gender age access to healthcare
2 diabetes mellitus chronic kidney disease corticosteroid use
3 glibenclamide metformin insulin admission
antimicrobials
4 Melioidosis
5 bacteraemia respiratory failure hypotension
6 Death
Figure 15. Conceptual hierarchical framework of risk factors for
melioidosis mortality.
Note.—Gender, age and access to healthcare occupy the highest level in the hierarchy
because they are not dependent on any other factors. Factors in each level are dependent
on factors in the level above and factors in lower levels cannot confound the effect of
factors in higher levels because they occur later in time. Factors in level 5 are immediate
proximate causes of death. Other pre-existing medical conditions were omitted from
level 2 (e.g., malignancy, thalassaemia, etc.) because they do not confound diabetes or its
treatment. We used occupation (rice farming) as a proxy for access to healthcare.
Measures of disease severity must also lie on the causal pathway. Why
then is it possible to adjust for drug therapy and not disease severity, when both
must lie on the causal pathway? The answer is that any exposure which changes
mortality must automatically change severity. It is implausible that any effective
intervention might reduce mortality while having no effect on clinical severity.
To adjust for severity in the analysis is an uninterpretable exercise, because the
clinician cannot modify severity independent of mortality.
2.2 Methods
We prospectively identified all patients aged 15 years or more presenting
to Sappasithiprasong Hospital, Ubon Ratchathani, northeast Thailand with
culture-confirmed melioidosis between 1 st January 2002 and 31 st December
2006. Patients on their first admission for culture-confirmed melioidosis were
82
Flow of time

eligible, but patients under the age of 15 were excluded because paediatric cases
have a different clinical presentation and prognosis. There were no other
exclusion criteria.
Patients were classified into three groups according to diabetes status at
presentation: known diabetes, hyperglycaemia or no diabetes. Patients with a
pre-existing diagnosis of diabetes mellitus were classified as having known
diabetes. The hyperglycaemia group were patients not previously known to have
diabetes, who had either a blood glucose >200 mg/dl (11·1 mmol/l) at any point
during the admission or new diabetes diagnosed after recovery as defined by
WHO criteria. We did not sub-classify this group further for the following
reasons: melioidosis is a common first presentation of type 2 diabetes in this
region 175 and an associated mortality of 50% (half of which occurs in the first 48
hours) means that a new diagnosis of diabetes cannot be made for a significant
proportion of cases, and hyperglycaemic patients who die are more likely to be
classed as having sepsis-induced hyperglycaemia while cases who survive are
more likely to be classed as having a new diagnosis of diabetes. Patients who did
not fall into the known diabetes or hyperglycaemia groups were classified as
having no diabetes. Hb A1c values were not available for patients in this cohort.
A full history and clinical examination were recorded from culture-
confirmed cases of melioidosis, and each patient was seen daily until discharge
or death. Blood samples were taken on admission for a complete blood count,
capillary glucose, blood urea nitrogen, creatinine, electrolytes and liver function
tests. Intravenous ceftazidime, imipenem or meropenem was started as soon as
melioidosis was suspected and continued for a minimum of 10–14 days. Clinical
data were recorded on a password-protected computer database.
Details of pre-admission drug treatment for diabetes (glibenclamide,
metformin and insulin) and duration of infection-related symptoms prior to
admission were obtained by questioning the patient or nearest relative. We
recorded whether effective admission antimicrobial therapy (parenteral
ceftazidime, imipenem, meropenem, co-amoxiclav, or cefoperazone-sulbactam)
was given within 24 hours of admission.
Sepsis was defined, using a simplified version of the 1992 joint American
College of Chest Physicians and Society of Critical Care Medicine Consensus
Conference Committee definition, as two or more of (1) temperature

38°C; (2) heart rate >90 beats per minute; (3) respiratory rate >20 breaths per
minute; or (4) total white cell count >12 × 10 9 cells/l. Bacteraemia was defined
as one or more blood cultures positive for B. pseudomallei. The extent of
infection was based on the number of organs or anatomical sites involved in the
infective process: Single organ disease meant involvement of a single site, and
multiorgan disease meant infection of more than one site. Blood culture and
throat swab positivity were excluded from the organ count. Pneumonia was
defined as the presence of clinical features consistent with pneumonia, plus
radiographic changes or sputum culture positive for B. pseudomallei. Duration
of symptoms prior to admission was recorded in days (but entered into the
analysis as fractions of weeks). A history was taken for chronic kidney disease,
thalassemia, malignancy and chronic liver disease (chronic hepatitis B or C
infection, or chronic alcoholic liver disease). Patients with nephrolithiasis were
identified from either the history or calculi visible on a plain abdominal
radiograph or ultrasonography.
The primary study outcome was in-hospital mortality, but we also pre-
defined two secondary outcomes: hypotension (a systolic blood pressure of less
than 90 mmHg at any point during admission), and respiratory failure (hypoxia
judged clinically to require mechanical ventilation; arterial blood gases are not
taken routinely in our setting).
2.2.1 Statistical analysis
Analyses were performed using Stata/SE 9 (StataCorp, College Station,
Texas). Differences between the three patient groups were compared using the
Fisher’s exact test for categorical variables and the Mann-Whitney U test for
continuous variables. Time to death (to a maximum of 28 days) was analysed
using the Kaplan-Meier method; patients discharged alive from hospital within
28 days were assumed to have survived, but patients who self-discharged
against medical advice were censored on the day of discharge.
To inform the selection of parameters used in the logistic regression
models, we created a conceptual hierarchical framework 700 of risk factors to
explore interactions between a number of variables (Figure 15). Parameters
were chosen on the basis of whether they were possible confounders for the
84

effect of diabetes on mortality (Model A, Table 10); 90 p-values reported are for
the likelihood-ratio test, but were not used to inform parameter selection.
In Model A (Table 10), we adjusted for age and sex as possible
confounders. We also adjusted for corticosteroid use because corticosteroids are
available without a prescription in northeast Thailand, are commonly abused,
and may cause diabetes. It is possible that patients in the known diabetes group
had increased access to healthcare, so we adjusted for this by using the
occupation of rice farmer as a proxy since this is associated with poverty,
increased physical distance from hospitals and clinics and poorer transport
links. It is impossible for parameters occurring below diabetes in the framework
to confound the effect of diabetes (e.g., treatment, pneumonia, severity, organ
failure). 700
A second regression model was developed to explore explanations for the
survival advantage in known diabetics (Model B, Table 10) by examining
modifiable parameters lying on the causal pathway between diabetes and
mortality. This included analysis of treatment for diabetes prior to admission,
and administration of appropriate admission antimicrobial therapy within 24
hours of admission. We adjusted for chronic kidney disease in this model
because metformin is contra-indicated in renal failure, and renal failure is
associated with mortality. Additional models were developed to evaluate the
effect of diabetes on our secondary outcomes (hypotension and respiratory
failure: Table 11, Models C & E), and to explore factors that could explain the
difference in outcomes observed in patients with diabetes (Table 11, Models D
& F).
The result of the Hosmer-Lemeshow test for goodness-of-fit for Model A
was p=0·41, B was p=0·53; C was p=0·22; D was p=0·28; E was p=0·51; and F
was p=0·47.
2.2.2 Glibenclamide inhibition study
Plates of Müller-Hinton agar were inoculated with a lawn of
B. pseudomallei (0·5 MacFarland standard). A single glibenclamide tablet was
crushed and dissolved in 1 ml DMSO, with 100 µl of the solution (~100 µg of
glibenclamide) dropped onto the centre of the agar plate. The same volume of
85

DMSO was used for control plates. The plates were incubated at 37°C for
48 hours and the plates examined for a zone of inhibition.
2.3 Results
2.3.1 Diabetes and mortality
We identified 1384 patients with culture-positive melioidosis, of whom
224 were removed from the final analysis for reasons shown in Figure 16. Of the
remaining 1160 patients, 410 (35%) had known diabetes, 250 (22%) had
hyperglycaemia and 500 (43%) had no diabetes. Patient characteristics, clinical
features of melioidosis and primary and secondary outcomes were compared
between the three groups, using the no diabetes group as the comparator for the
two other groups (Table 9). In-hospital mortality was lower in patients with
known diabetes compared with non-diabetics (p=0·04), with no difference
observed between non-diabetics and patients with hyperglycaemia. These
findings were reproduced in the survival analysis (Figure 17A).
500 (43%)
No diabetes
1384 culture-positive
1182 eligible
1160 study patients
410 (35%)
Known diabetes
Figure 16. Summary of patient recruitment for cohort study.
86
202 excluded
• 129 age

No diabetes Known diabetes
Hyperglycaemia
Total
(n=500) (n=410)
(n=250)
(n=1160)
number (%) number (%) p-value a number (%) p-value a row total (%)
Female 165 (33) 219 (53) – 114 (46) – 498 (43)
Male 335 (67) 191 (47)

Distribution of disease
Bacteraemia 274 (55) 238 (58) 0.35 164 (66) 0.005 676 (58)
Single organ disease 291 (58) 240 (59) 0.95 146 (58) 1.00 677 (58)
Multiorgan disease
Complications
110 (22) 102 (25) 0.31 58 (23) 0.71 270 (23)
Hypotension 187 (37) 135 (33) 0.16 99 (39) 0.58 421 (36)
Respiratory failure 179 (36) 117 (29) 0.02 101 (40) 0.23 397 (34)
Sepsis d 392 (78) 334 (81) 0.28 209 (84) 0.10 935 (81)
Antibiotic treatment and in-hospital mortality
Effective antibiotic treatment
within 24 hours of admission
367 (73) 359 (88) 20 breaths per minute, or total leucocyte count
>12 × 10 9 cells/L.
f Patients who took their own discharge were counted as alive.
88

Figure 17A
Survival
0.00 0.20 0.40 0.60 0.80 1.00
0 7 14 21 28
Survival time in days
Numbers at risk on day 0 day 7 day 14 day 21 day 28
No diabetes 500 304 281 271 265
Known diabetes 410 279 259 252 251
Hyperglycaemia 250 154 138 130 130
—— No diabetes —— Known diabetes —— Hyperglycaemia
Known diabetes v. No diabetes (p = 0·03)
Hyperglycaemia v. No diabetes (p = 0·91)
Known diabetes v. Hyperglycaemia (p = 0·05)
89

Figure 17B
Survival
0.00 0.20 0.40 0.60 0.80 1.00
Numbers at risk on day 0 day 7 day 14 day 21 day 28
No diabetes 500 304 281 271 265
Known diabetics on glibenclamide 209 158 153 153 149
Known diabetics not on glibenclamide 201 121 106 103 102
Hyperglycaemia 250 154 138 130 130
—— No diabetes group —— Known diabetes on glibenclamide
—— Known diabetes not on glibenclamide —— Hyperglycaemia group
Known diabetes on glibenclamide v. No diabetes (p

A previous study conducted at the same hospital but in a different (earlier)
patient cohort reported an inverse association between diabetes and
mortality (unadjusted odds ratio [OR] 0·49). 681 We confirmed this
observation in our independent cohort (OR 0·76; adjusted odds ratio
[AOR] 0·78, 95% confidence interval [CI] 0·59 to 1·0, p=0·07), and
extended previous observations that patients in the hyperglycaemia group
had no survival advantage (Model A, Table 10).
2.3.2 Glibenclamide and mortality
We hypothesised that the association between diabetes and a reduced
risk of death was due to the treatment received for diabetes and/or
melioidosis. This was tested in a second logistic regression model (Table
10, Model B). We considered drugs prescribed for diabetes (glibenclamide,
metformin, insulin) prior to admission, noting that more than half of
known diabetics were receiving glibenclamide (Table 9). We adjusted for
chronic renal failure, because metformin is avoided in renal failure,
because of the risk of lactic acidosis, and because renal failure is associated
with mortality. We included the administration of appropriate admission
antimicrobial therapy because the association between diabetes and
melioidosis is well-described and could lead clinicians to prescribe these at
an earlier stage in diabetic patients.
The adjusted odds ratio for death was lower for patients who received
glibenclamide therapy prior to admission (AOR 0·47, 95%CI 0·28–0·74,
p=0·005) and for those receiving effective antimicrobial chemotherapy at
admission (AOR 0·23, 95%CI 0·17–0·34, p

univariable analysis logistic regression
model A
92
(n = 1160)
logistic regression
model B
(n = 1109) c
OR (95% CI) AOR (95% CI) AOR (95% CI)
No diabetes a 1·0 — 1·0 — 1·0 —
Known diabetes 0·76 (0·58–0·99) 0·78 (0·59–1·0) 1·4 (0·89–2·3)
Hyperglycaemia 1·1 (0·79–1·5) 1·1 (0·81–1·5) 1·4 (0·98–1·9)
Age b 1·1 (1·0–1·2) 1·1 (1·0–1·2) 1·1 (0·99–1·2)
Male sex 1·2 (0·94–1·5) 1·1 (0·89–1·5) 1·0 (0·78–1·3)
Rice farming 0·97 (0·74–1·3) 1·0 (0·75–1·3) 1·2 (0·86–1·6)
Corticosteroid use 1·5 (0·86–2·7) 1·6 (0·87–2·8) 1·0 (0·62–2·2)
Chronic kidney disease 2·8 (1·8–4·3) 2·2 (1·6–4·0)
Glibenclamide treatment c 0·48 (0·35–0·67) 0·47 (0·28–0·74)
Metformin treatment c 0·61 (0·33–1·1) 1·1 (0·62–2·4)
Insulin treatment c 0·79 (0·49–1·3) 0·65 (0·37–1·2)
Effective admission antibiotic
treatment c
0·22 (0·16–0·31) 0·23 (0·17–0·34)
Table 10. The effect of diabetes on mortality.
Note.—OR = odds ratio (not adjusted); AOR = adjusted odds ratio; CI = confidence interval. An odds ratio below 1 indicates association with
survival, whereas an odds ratio above 1 indicates association with mortality. The first column describes the contribution of each factor in isolation.
The second column (model A) attempts to explain the effect of diabetes by adjusting for several possible confounders for diabetes simultaneously;
the third column (model B) adjusts additionally for the effect of diabetes treatment and post-admission antibiotics.
When considered in isolation, diabetes (OR 0·76), glibenclamide treatment (OR 0·48) and effective admission antibiotic treatment (OR 0.22) were
associated with survival. The effect of diabetes persisted after correcting for confounders for diabetes (AOR 0·78, model A). When the effect of
glibenclamide treatment and effective admission antibiotic were taken into account, diabetes was no longer associated with survival (AOR 1·4,
model B). Sensitivity analysis. We conducted a sensitivity analysis to examine the impact of the missing data. Model A was constructed by

assigning the patients who self-discharged as “alive”; but if these patients were instead assigned as “dead”, then the AOR for known diabetes in
model A became 0·71 (0·54–0·93, p = 0·01). In model B, assigning the patients who self-discharged to “dead”, caused the AOR for known diabetes
to become 1·3 (0·79–2·1, p = 0·31), glibenclamide treatment AOR 0·48 (0·28–0·83, p = 0·008), and effective admission antibiotic treatment AOR
0·18 (0·13–0·27, p

Table 11
Hypotension
univariable analysis logistic regression
Model C
(n = 1160)
94
logistic regression
Model D
(n = 1109) a
OR b (95% CI c ) AOR d (95% CI c ) AOR d (95% CI c )
No diabetes e 1·0 — 1·0 — 1·0 —
Known diabetes 0·80 (0·62–1·0) 0·84 (0·63–1·1) 1·3 (0·79–2·0)
Hyperglycaemia 1·2 (0·90–1·6) 1·1 (0·83–1·6) 1.3 (0·91–1·7)
Age f 1·2 (1·1–1·3) 1·1 (1·0–1·3) 1·1 (1·0–1·2)
Male sex 1·1 (0·87–1·4) 1·1 (0·87–1·4) 1·0 (0·77–1·3)
Rice farming 1·0 (0·75–1·4) 1·0 (0·76–1·4) 1·1 (0·78–1·5)
Corticosteroid use 2·9 (1·6–5·2) 3·0 (1·6–5·4) 2·4 (1·3–4·5)
Chronic kidney disease 2·0 (1·3–3·0) 1·8 (0·92–3·5)
Glibenclamide treatment a 0·56 (0·40–0·78) 0·48 (0·30–0·78)
Metformin treatment a 1·1 (0·60–1·9) 1·8 (0·92–3·5)
Insulin treatment a 0·69 (0·42–1·2) 0·60 (0·33–1·1)
Effective admission
antimicrobial therapy a
0·46 (0·34–0·62) 0·51 (0·37–0·70)

Respiratory failure
univariable analysis logistic regression
Model E
(n = 1160)
95
logistic regression
Model F
(n = 1109) a
OR b (95% CI c ) AOR d (95% CI c ) AOR d (95% CI c )
No diabetes e 1·0 — 1·0 — 1·0 —
Known diabetes 0·72 (0·54–0·95) 0·73 (0·54–0·97) 1·2 (0·74–1·9)
Hyperglycaemia 1·2 (0·89–1·7) 1·2 (0·90–1·7) 1·4 (1·0–2·0)
Age f 1·0 (0·94–1·1) 1·0 (0·93–1·1) 0·99 (0·92–1·1)
Male sex 1·2 (0·94–1·5) 1·2 (0·91–1·5) 1·1 (0·82–1·4)
Rice farming 0·98 (0·73–1·3) 0·99 (0·73–1·3) 1·0 (0·76–1·4)
Corticosteroid use 1·6 (0·90–2·8) 1·7 (0·92–3·0) 1·2 (0·62–2·2)
Chronic kidney disease 2·2 (1·5–3·3) 2·0 (1·2–3·1)
Glibenclamide treatment a 0·44 (0·31–0·64) 0·50 (0·28–0·86)
Metformin treatment a 0·52 (0·26–1·0) 1·0 (0·48–2·1)
Insulin treatment a 0·80 (0·49–1·3) 0·75 (0·39–1·4)
Effective admission
antimicrobial therapy a
0·40 (0·29–0·54) 0·42 (0·31–0·58)
Table 11. Effect of diabetes on hypotension and respiratory failure.
Note.—
a Patients on unknown oral diabetes medication were omitted (n = 1109) in Models D & F and in the unadjusted OR for treatment variables.
b Odds ratio (unadjusted).
c Confidence interval.
d Adjusted odds ratio.
e Comparator group.
f Number of decades above age 15 years.

Patients in the known diabetes group were more likely to receive effective
antimicrobial therapy on admission compared to patients without diabetes
(88% versus 73%, p

these two factors, but none of the previously-cited studies 574,594,595 looked for an
effect of diabetes treatment.
The effect of antimicrobial chemotherapy is unsurprising, and is consistent
with what we know from other causes of sepsis. 701 The effect of glibenclamide
was unexpected and has not previously been described.
2.4.1 Glibenclamide
Glibenclamide rINN (=glyburide USAN, Figure 18) is a second-generation
sulphonylurea widely used to treat type 2 diabetes and acts by inhibiting ATP-
sensitive 699 (KATP-channels) in pancreatic β-cells, 95 leading to stimulation of
insulin secretion.
Figure 18. Glibenclamide rINN or glyburide USAN.
Note.— IUPAC name, N-[2-[4-(cyclohexylcarbamoylsulfamoyl)phenyl]ethyl]-2-
methoxybenzamide. Molecular weight 494·0 g/mol. Daonil® (formerly Hoechst, now
Aventis); Eugluclon® (Roussel); Sigma-Aldrich catalogue no. G0639.
A B
Figure 19. General structure of sulphonylureas and sulphonamides.
Note.—H = hydrogen; N = nitrogen; O = oxygen; R = general alkyl group, S = sulphur.
The generalised structure on the left is a sulphonylurea (A); the structure on the right is a
sulphonamide (B). The part of the structure highlighted in red is the sulphonamide
moiety, which is clearly present in the sulphonylurea structure as well.
97

The sulphonylureas were first described in the 1942 by Janbon et al. while
searching for novel antimicrobials. From a chemical point of view, the
sulphonylureas are also sulphonamides (Figure 19) and we initially
hypothesised that glibenclamide was acting via an antimicrobial action on
B. pseudomallei. However, we were unable to show that glibenclamide inhibits
growth of B. pseudomallei, thus excluding an antimicrobial effect of
glibenclamide on B. pseudomallei.
2.4.2 Glibenclamide and vascular smooth muscle
KATP-channels are also expressed in the heart and on vascular smooth
muscle, and excessive opening of vascular KATP-channels in sepsis (due to a fall
in the ATP/ADP ratio, pH or oxygen tension, or a rise in intracellular lactate)
leads to hyperpolarization of the cell membrane, a blocking of the calcium influx
and loss of vascular tone (contributing to the ‘vascular paresis’ of septic shock).
By blocking vascular smooth muscle KATP-channels, glibenclamide is able to
protect from shock in animal models of sepsis, 702 although this did not translate
into a detectable effect on cardiovascular parameters in two small clinical trials
in which glibenclamide was given at doses normally used to treat diabetes. 703,704
Neither study was designed to find an effect on mortality, and neither looked at
end points such as respiratory failure or inflammation.
2.4.3 Limitations of this study
In seeking confounders for our study, we considered whether
glibenclamide treatment was simply a marker for less advanced diabetes and
that this was associated with a survival benefit compared with patients on
insulin which is reserved for patients who fail to achieve adequate diabetic
control on oral medication. However, only glibenclamide treatment and not
metformin was associated with an improvement in mortality. We also
considered whether thiazolidinediones and statins (which are anti-
inflammatory and may be required more commonly in diabetics) was a
confounder, but found that although available, these are very rarely used in our
setting. This is an observational study in which glibenclamide was given for a
specific indication, namely, the treatment of type 2 diabetes and it is not
possible to exclude unequivocally the possibility of confounding. 705
98

3 Glibenclamide is associated with
reduced inflammation in melioidosis
3.1 Introduction
Having observed an association between glibenclamide and survival, we
initially postulated that glibenclamide was acting as an antimicrobial drug based
on its structural homology with the sulphonamides (to which B. pseudomallei is
susceptible). However, a solution of glibenclamide in DMSO has no inhibitory
effect B. pseudomallei grown in LB or on tryptic soy agar. We therefore
hypothesized that glibenclamide was having an effect on the host response to
melioidosis.
Glibenclamide is a broad-spectrum ABC transporter inhibitor that alters
responses of macrophages to a range of stimuli in vitro and in vivo. 706–708 We
used peripheral leukocyte gene expression as a screening tool to identify global
trends in gene function that might be ascribed to the action of glibenclamide.
We compared gene expression profiles in 10 diabetic patients who were taking
glibenclamide at the time of presentation with melioidosis against 10 diabetic
melioidosis patients who were not taking this or another sulfonylurea, and
repeated the comparison in healthy controls.
3.1.1 DNA microarrays
DNA microarrays are a multiplex assay, consisting of two-dimensional
arrays of DNA fragments immobilized on a solid substrate. The first DNA
microarray had 378 probes, but current commercial arrays may contain tens of
thousands of probes and are able to interrogate across an entire eukaryotic
genome. When investigating gene expression, messenger RNA is first purified
from lysed white blood cells and then converted to cDNA. The cDNA is then
amplified by transcription to RNA by an RNA polymerase, which is preferable to
PCR amplification because the amplification proceeds in a linear fashion that
preserves the relative proportions of each mRNA transcript. A fluorescent label
is attached to the RNA, and the RNA is then hybridised with a DNA microarray,
where each RNA fragment will bind to matching DNA probes on the chip. The
99

fluorescence read out gives an estimate of the amount of RNA bound to each
probe.
Microarrays are a refinement of the Southern blot. DNA blotting was first
described by Edwin Southern in 1975 at the University of Edinburgh. 709
Southern described a method for identifying specific sequences by separating
DNA fragments electrophoretically by size on an agarose gel, blotting the DNA
onto nitrocellulose filters, then detecting the sequences of interest using
complementary radiolabelled or fluorescent-labelled DNA probes. 709
Microarrays arose from the need for an assay that was able to
simultaneously detect the presence or absence of multiple sequences. In 1982,
at the Sloan-Kettering Institute in New York, Augenlicht and Kobrin realised
that it was not necessary to separate the DNA fragments in the sample, but that
similar results could be obtained by spatially separating and immobilising the
probes in an array instead. Their first array consisted of 378 probes immobilised
on nitrocellulose membranes, on which they were able to semi-quantitatively
describe differences in gene expression in a mouse tumour. 710 Five years later,
the same team published a protocol for computerised scanning and image
processing that allowed quantitative analysis of 4000 human sequences
simultaneously. 711
Affymetrix was founded by Stephen Fodor in 1992 and has its
headquarters in Santa Clara, California. Affymetrix began marketing the first
commercially viable DNA microarrays (under the trademark, GeneChip®) in
1994, produced using semiconductor manufacturing techniques, and this made
microarray technology widely available to researchers.
3.1.2 Illumina BeadChip
Illumina was founded in 1998 to develop and manufacture integrated
systems for the analysis of genetic variation of function. The company was
founded to exploit the BeadArray technology developed at Tufts University and
has its headquarters in San Diego, California. The first BeadArray application
was to genotype single-nucleotide polymorphisms, but has expanded to include
copy number variation analyses, DNA methylation and gene expression
profiling.
100

In a BeadArray, each microarray consists of hundreds of thousands of
3 µm silica beads randomly assembled in 3 µm microwells on silica glass slides,
and each silica bead is coated with hundreds of thousands of copies of a single
complementary DNA (cDNA) probe. The arrays are self-assembling, with each
bead bound to its well by van der Waals forces alone.
A B
Figure 20. Illumina BeadArray.
Notes.—A Illumina BeadArrays are distinguished from other microarrays principally by
their method of manufacture. Conventional microarrays are created by printing spots
onto slides or membranes: BeadArrays are created from randomly self-assembling beads.
First, 3 µm microwells are etched into glass slides, then a mixture of glass beads, each
coated with hundreds of copies of a single probe are poured onto the slide. The 3 µm
beads slot randomly into wells and are held there by van der Waals forces. Excess beads
are then washed off. Each probe is represented by ~30 beads distributed randomly across
the array. B The glass slide on the right is the Illumina HumanWG-6 v.3 BeadChip used
for this project. Each BeadChip contains twelve strips (two strips per array, one sample
per array). In contrast, Affymetrix GeneChips will only take a single sample per chip.
Images taken from Illumina product literature.
Each probe is replicated by ~30 beads per strip (the actual number of
beads varies randomly from ~15 to ~60) and the final position of each bead is
completely random. Each probe consists of a 50-mer probe sequence and a 29-
mer address sequence. The address sequence is decoded at the factory to reveal
101

the probe sequence at each position. Each chip is shipped with this information
and this information is unique for each chip. The sequence of these address tags
is commercial confidential.
This method of manufacture has a number of advantages. Conventional
arrays are printed onto slides or membranes using pins. Small defects in
individual pins result in systematic errors that propagate across the entire array,
because every spots printed by the same pin will carry the same defect. The
BeadArray technology relies on random self-assembly of the array, which means
that this kind of systematic error cannot occur. A further advantage of the
BeadArray is the high degree of replication, which means that Illumina
BeadArrays are robust to spatial defects. As much as three-quarters of each strip
could be destroyed with only minimal loss of information. By contrast,
Affymetrix GeneChips each contain only one copy of each probe at a constant
location on every chip: GeneChips are therefore extremely susceptible to spatial
artefacts, with no way of recovering from these defects. A final advantage is that
BeadArrays are cheaper to manufacture than conventional spotted arrays and
cheaper to use. At the time the project was planned. Affymetrix GeneChips
would only take a single sample per GeneChip, whereas a BeadChip could take 6
or 8 samples per chip.
The microarray used in this project was the HumanWG-6 v3.0, 712 which
contains 12 strips per glass slide, or six samples (two strips per array, one array
per sample). The array covers 48,804 probes or 25,440 manually curated genes
from the NCBI RefSeq 713 (build 36.2, release 22) and UniGene 714 databases
(build 199). The genes included cover the whole human genome and are not
restricted to well-characterised genes, but also includes gene candidates and
splice variants. 712 The platform is single colour (Cy3, which fluoresces green),
which avoids the problems associated with ozone degradation from which two-
colour platforms suffer, 715 and the need for complicated experimental designs
involving dye swaps.
3.1.3 Analysis of microarray data
The output of a microarray experiment consists of a fluorescence value per
probe per sample. Fluorescence values are proportional to absolute expression
values, but are measured in arbitrary units. They therefore gain meaning only in
102

comparisons between groups (treated versus untreated, wild-type versus
knockout, etc.). This also means that it is very hard to compare raw data
acquired by different scanners. The most commonly used statistical test is the
Student t-test (or a modified version thereof, such as a moderated t-test or B-
statistic).
The primary difficulty in the statistical analysis of microarray results is the
large number of statistical tests that are carried out, often with small sample
sizes. Should one set a p-value cut-off of 5% then one expects 5% of the tests to
have a p-value

conservative method than family-wise error correction methods (which are the
Bonferroni method and its relatives), but in essence, the method also works by
changing the p-value threshold to achieve the desired false discovery rate.
The major criticism of these methods this is that they all simultaneously
increase the type 2 error rate while reducing the type 1 error rate. 720 A type 2
error occurs when a true difference is missed by a statistical test (i.e., a false
negative) and it has been argued that type 2 errors are as important as type 1
errors. 720 The assumptions underlying these methods have also been
questioned: if each test is independent, then why should the result of each test
be corrected for the number of tests performed? The number of tests performed
depends on the investigator and is not a characteristic of the test, so it seem
irrational to address the problem by modifying the test 720 (in slightly more
technical terms: these are mathematically rigorous methods for maintaining the
family-wise error rate, but there are no strong arguments why maintaining the
family-wise error rate should be desirable in the first place). It should also be
noted that all p-value cut-offs are arbitrary (including the ‘conventional’
threshold of 0·05), because all p-values exist on a continuum from 0 to 1, and
therefore the decision to set any threshold (whether 0·001 or 0·1) may be
criticised. 721 It is also worth mentioning that even in classical significance
testing, the decision is not merely to accept or reject the null hypothesis. Keppel
has pointed out that under certain circumstances, a third option may be more
appropriate, which is to suspend judgement. 722
The strongest single argument against adjusting for multiple comparisons
is that blind reliance on stringent statistical criteria gives results that cannot be
reproduced. 723 The MicroArray Quality Control project (MAQC) is an initiative
of the US Food & Drug Administration, the aim of which is to assess the
reproducibility of microarray experiments and the comparability of results
across different platforms. In a study published in 2006, two distinct, high-
quality RNA samples were assayed at 137 different sites on seven different
microarray platforms across the continental USA. A major finding of that study
was that reproducibility falls as statistical stringency increases, leading the
investigators to recommend that non-stringent p-value cutoffs be used for
microarray experiments. 723
104

A second (complementary) method of dealing with the gene lists generated
by microarray experiments is to group them by function. The rationale behind
this is that genes that are truly differentially regulated are likely to share the
same function. In contrast, the false positives in the gene list are likely to come
from any part of the genome and are less likely to form functional groupings. 724–
726 This group of methods is commonly known as gene set enrichment and is a
necessary part of the interpretation of any gene set.
Broadly speaking, there are two main functional groupings: the first is
pathway and the second is ontology.
Pathways may be defined most generally as a sequence of biochemical
interactions within a cell. Metabolic pathways are easy to define and
circumscribe, because they follow the serial modifications of a single molecule,
where the product of one reaction is the substrate for another. For example, in
glycolysis (the first metabolic pathway to be described), glucose (substrate) is
phosphorylated to glucose-6-phosphate (product) by the enzyme, hexokinase.
Glucose-6-phosphate is the substrate for phosphoglucose isomerase, and so
forth down the pathway, ending with pyruvate or acetyl-CoA. In signal
transduction, there is not usually a single molecule that can be followed, but
instead the trail follows a notional ‘signal’ that is passed from protein to protein.
The first signalling pathway to be described was glucagon signalling by Martin
Rodbell (who first described G-proteins, for which he won the 1994 Nobel Prize
for Medicine). 727,728 Signalling pathways follow a series of activation steps and
secondary messengers. Unlike metabolic pathways, there is usually no single
molecule that follows the course of the entire pathway. Although signalling
pathways are usually defined as the events downstream of an initiating signal
molecule (e.g., insulin or adrenalin), that molecule seldom progresses along the
entire pathway. Pathways often interact, multiple signalling molecules may
initiate identical events in the same pathway, and some pathways may require
multiple signals. The ‘canonical’ or textbook description of any signalling
pathway is therefore a subjective construct, its boundaries defined by ease of
interpretation and not by biological function. The descriptions of many
pathways will therefore vary from textbook to textbook and from database to
database, with no clear reasons why some proteins are included and some
omitted. The pathways for coagulation and inflammation have multiple
105

initiators and are even harder to describe clearly or consistently because of
multiple overlaps and interactions, with some molecules having apparently
contradictory roles (for example, IL10 is said to be both pro-inflammatory and
anti-inflammatory). There are two currently two large systematic efforts to
develop canonical databases of known pathways: Reactome 729 and KEGG (the
Kyoto Encyclopedia of Genes and Genomes). 730
Gene ontology is an even more nebulous concept and the dictionary
definition is ‘the branch of metaphysics dealing with the nature of being’. A
more helpful definition is that gene ontology is accurate and succinct
description of genes. There are a number of projects that do this, the largest of
which is the Gene Ontology project. 731 The GO project is developing and
maintains three separate ontologies (controlled, structured vocabularies) for the
description of gene products. These are (i) cellular component (e.g., nucleus or
membrane or organelle), (ii) biological process (e.g., inflammation or small
molecule transport or DNA replication), and (iii) molecular function (e.g.,
protein kinase or transmembrane transporter or oxidoreductase).
We elected to use Innate DB for our analysis because of its focus on
manually curating pathways related to innate immunity. 725 It is also linked to
other protein interaction databases and uses standard nomenclature such as
those developed by HUGO and GO. The database takes lists of genes and
clusters them by pathway or by ontology, the statistical test being the
hypergeometric test.
The techniques for analysing microarray data (and high throughput data
generally) are still an active area for research. There is therefore no definitive
‘best practice’ defined, but the literature support an approach that uses a
relatively lax statistical criteria, so that both type 1 and type 2 error rates are
reasonably controlled, then use functional analysis to eliminate false positives
and aid interpretation.
The main barrier to exchanging microarray data and to reproducing
microarray experiments is the complete lack of standardisation of fabrication,
assay protocols or analysis methods. Many of these details of fabrication are
commercial confidential, and assay protocols are usually specific to platform.
However, in theory, analyses should be reproducible. The MIAME checklist
(minimum information about a microarray experiment) defines the minimum
106

level of detail that should be reported for all microarray experiments, and has
been adopted by many journals as a requirement for the submission of papers
that incorporate microarray results, 732 with the aim of improving the
reproducibility of microarray experiments. 733
3.2 Methods
3.2.1 Study subjects
We compared peripheral white blood cell gene expression in (i) diabetics
who were taking glibenclamide (case) or were not taking this or another
sulfonylurea (control) at the time of admission to Sappasithiprasong Hospital
with culture-confirmed melioidosis and sepsis, and (ii) otherwise healthy
diabetics attending a routine out-patient clinic who were taking glibenclamide
(case) or were not taking this or another sulfonylurea (control). Eligible cases
for both studies were persons aged between 18 and 75 years. A total of 40
patients (10 in each group) were recruited in the period 31 Jan 2008 to 31 Oct
2008. Diabetes was defined as an abnormal Hb A1c at enrolment (7.8% or
greater, 734 Bio-Rad® D-10 fully-automated Hb A1c system, Bio-Rad
Laboratories) or a previous diagnosis of diabetes. The Hb A1c concentration
allowed us to identify patients with previously unrecognized diabetes who went
on to die before we could make a diagnosis of diabetes by WHO criteria.
We obtained approval from the Oxford Tropical Research Ethics
Committee and the Ethics Committee of the Faculty of Tropical Medicine,
Mahidol University for the gene expression study. Written informed consent
was obtained from all subjects by a native Thai speaker. All procedures
performed were in accordance with the Helsinki Declaration of 1975 (revised
1983).
3.2.2 RNA extraction and microarray analysis
A 3 ml blood sample was collected from each subject in a PaxGene TM Blood
RNA tube (PreAnalytiX) and stored at –70°C prior to mRNA extraction
according to the manufacturer’s instructions. At study completion, blood
samples collected in PaxGene TM Blood RNA tubes were transported in one
shipment on dry ice to the Wellcome Trust Sanger Institute, Cambridge, where
107

RNA was extracted using the PaxGene TM Blood RNA Purification Kit
(PreAnalytix, GmbH) according to the manufacturer’s instructions.
RNA purity was measured using the ratio of absorbance at 260 and 280
nm and all samples had ratios between 1·7 and 2·1. RNA integrity was evaluated
using the Bioanalyzer (Agilent) and all samples had RNA integrity numbers
greater than 6. The RNA was amplified using the Illumina® TotalPrep RNA
Amplification Kit (Applied Biosystems, Carlsbad, California) and assayed using
the Illumina® HumanWG-6 v3.0 Expression BeadChip (Illumina, Cambridge,
UK), which probes 48,804 transcripts from across the human genome. We
examined all genes with ≥1·5-fold difference in expression with the false-
discovery rate set to 10% (Student’s t-test with Benjamini-Hochberg
correction; 719 GeneSpring GX version 10.0, Agilent Technologies, California).
3.2.3 Quantitative reverse-transcriptase polymerase chain reaction
Results were confirmed by quantitative reverse-transcriptase polymerase
chain reaction (qPCR): 500 ng of sample RNA was reverse transcribed using the
QuantiTect Reverse Transcription kit (Qiagen 205313) according to the
manufacturer’s instructions. Primers were ordered from Qiagen gene globe, as
follows: IL18R1 catalogue number QT00082922; IL8, QT00000322; TNF,
QT01079561; IFNG, QT00000525; ACTB1, QT00095431. Real time PCR was
performed in triplicate using the QuantiTect SYBR Green PCR kit (Qiagen
204145) according to the manufacturer’s instructions on a StepOne Plus
(Applied Biosystems) for 40 cycles. Melting curve analyses were performed for
all PCR products. The β-actin housekeeping gene was selected as loading
control because of the stability of its expression in the microarray analysis. Fold
change was calculated using the comparative Ct method.
Gene ontology over-representation analysis was performed using
InnateDB 725 (http://www.innatedb.ca/) and the p-values reported are for the
hypergeometric test. Gene lists were supplemented by manual literature
searches for genes not curated by InnateDB. Microarray data have been
deposited at ArrayExpress, EMBL-EBI (accession number E-TABM-852-n).
3.2.4 Neutrophil stimulation study
Neutrophils from three healthy male caucasian volunteers were isolated
using Polymorphprep (Axis-Shield, Oslo, Norway) according to the
108

manufacturer’s instructions, then resuspended (final concentration 1·0 × 10 6
cells/ml) in RPMI 1640 (Gibco, Invitrogen, Paisley, UK) + L-glutamine + 15%
heat-inactivated fetal bovine serum (Lonza BioWhittaker, Verviers, Belgium).
Purity was >99% by flow cytometry (Becton Dickinson FACSCalibur; monocytes
and neutrophils identified by their characteristic scatter profile and CD14
staining). Cells were incubated with medium or 1.0 mM glibenclamide
(comparable to the peak human plasma concentration achieved following a
20 mg oral dose) for 60 minutes in a 96-well plate, then stimulated with
medium or heat-killed B. pseudomallei 1026b (5 × 10 7 cfu/ml) for 4 hours at
37°C in 5% carbon dioxide. The cells were centrifuged (300 × g, 5 minutes) and
the supernatant removed. The pellet was resuspended (5% bovine serum
albumin w/v, 350 µM EDTA, 0·01% sodium azide w/v in phosphate-buffered
saline) and stained for CD14 and CD11b (BD Pharmingen, Erembodegem-Aalst,
Belgium). IL1β (Invitrogen, Camarillo, California), IL8 (Arcus Biologicals,
Modena, Italy) and TNFα (Arcus) were measured by enzyme-linked
immunosorbent assay according to the manufacturers’ instructions. Paired
t-tests were performed using Prism 5.0c for Mac (GraphPad Software, San
Diego, California).
3.3 Results
There were no significant differences in any of the clinical parameters
recorded for the two patient groups (Table 12). The expression of 205 probes
(representing 186 distinct genes), were significantly different in patients taking
glibenclamide compared to those who were not (Figure 21). Sixty-three
immune-related genes were differentially expressed in melioidosis patients on
glibenclamide (p=0·001), the net effect of which would be postulated to be anti-
inflammatory (Appendix A). Prominent among the immune-related genes
include those implicated in neutrophil activation, endothelial adhesion and
transmigration. There was no statistical difference in neutrophil, lymphocyte or
monocytes counts between the two groups (Table 12), indicating that this was
not merely an effect of differences in cellular blood constituents. In the
microarray analysis, tumor necrosis factor (TNF)-α, interleukin (IL)-18R1 and
IL-8 expression were all down-regulated in glibenclamide-treated patients;
interferon (IFN)-γ showed a modest trend to down-regulation. Quantitative
109

everse-transcriptase PCR corroborated the TNFα and IL18R1 results; IFNγ was
down-regulated by this methodology, but the IL8 results were not confirmed
(Table 13).
We sought further evidence for an effect of glibenclamide on gene
expression profiles in otherwise healthy diabetics who were taking
glibenclamide (n=10) or who were not taking a sulfonylurea drug (n=10). There
were no significant differences in baseline parameters between the two groups
(Table 12). We found that 31 probes (29 distinct genes) were significantly
different in patients taking glibenclamide compared to those not, 13 of which
were immune related (Appendix A), again suggesting an anti-inflammatory
effect. Of note, genes associated with neutrophil function were down-regulated
in the glibenclamide group (p

Diabetes patients with melioidosis
Parameter Taking glibenclamide Not taking
glibenclamide
p-value
Male gender 6 of 10 5 of 10 1·00 a
Age (years) 60±7 51±9 0·02 b
Glucose (mg dL –1 ) 218±64 256±111 0·37 b
Hb A1c (%) 10·8±3.2 11·1±3.3 0·83 c
Neutrophils (×10 9 L –1 ) 10·8±8·5 8·8±4.6 0·52 b
Lymphocytes (×10 9 L –1 ) 1·6±0.9 1·5±1.4 0·93 b
Monocytes (×10 9 L –1 ) 0·8±0·3 0·5±0·4 0·05 b
Eosinophils (×10 9 L –1 ) 0·1±0·1 0·1±0·1 0·82 c
Mortality 2 of 10 5 of 10§ 0·41 a
Diabetes patients who were otherwise healthy
Parameter Taking glibenclamide Not taking
glibenclamide
p-value
Male gender 3 of 10 5 of 10 0·65 a
Age (years) 54±11 56±11 0·66 b
Glucose (mg dL –1 ) 126±31 124±48 0·93 b
Hb A1c (%) 9·0±2·1 9·0±2·2 0·98 c
Neutrophils (×10 9 L –1 ) 4·7±1·3 5·1±2·1 0·63 c
Lymphocytes (×10 9 L –1 ) 4·4±1·4 5·1±2·1 0·40 b
Monocytes (×10 9 L –1 ) 0·4±0·2 0·6±0·3 0·19 b
Eosinophils (×10 9 L –1 ) 0·7±0·4 0·8±1·3 0·22 c
Table 12. Comparison of gene expression study patient
characteristics.
Note.—Values are reported as means, except where stated. Errors reported are standard
deviations.
a Fisher exact test.
b Welch t-test.
c Student t-test.
d One patient in this group was lost to follow-up following discharge from hospital, and
was counted as having survived to discharge.
111

Figure 21. Differential expression of inflammation-associated genes
in diabetic patients with acute melioidosis with or without
glibenclamide.
Note.—Gb = glibenclamide. Genes that are up-regulated are shown in red, down-
regulated in green; genes in black are not differentially expressed. The gene symbols used
are those assigned by the HUGO gene nomenclature committee.
112

Gene Fold change attributable to
glibenclamide therapy in melioidosis
patients
IFNγ –2·32
IL8 +4·23
IL18R1 –1·39
TNFα –15·2
Table 13. Results of quantitative reverse-transcriptase polymerase
chain reaction.
Note.— IFNγ= gamma interferon; IL8 = interleukin 8; IL18R1 = interleukin 18 receptor 1;
TNFα = tumour necrosis factor alpha.
Figure 22. Effect of glibenclamide pre-treatment on neutrophil
stimulation
Note.—Bps = heat-killed Burkholderia pseudomallei; Gb = glibenclamide,
IL = interleukin. The interrupted line marks the limit of detection. Black dots represent
control samples; open circles are samples pre-treated with glibenclamide. Isolated human
neutrophils (1·0 × 10 6 cells/ml, >99% purity) were treated for 60 minutes with 1·0 mM
glibenclamide then stimulated with heat-killed B. pseudomallei (5 × 10 7 cfu/ml). The cells
were incubated at 37°C for four hours in 5% carbon dioxide, then CD11b expression was
measured by flow cytometry. IL8 concentrations were measured in supernatant by
enzyme-linked immunosorbent assay. IL1β and TNFα concentrations were below the
limit of detection. The p-values reported are for paired t-tests.
113

3.4 Discussion
Glibenclamide is known to have a wide-range of anti-inflammatory effects,
one of which is an inhibitory effect on inflammasome assembly. 735 The
inflammasome is an intracellular protein complex present in macrophages that
activates caspase 1 and converts pre-formed IL1β and IL18 to their active forms
when presented with an appropriate inflammatory stimulus. 459 IL1β and IL18
levels are high in bronchoalveolar lavage fluid from patients with acute lung
injury/acute respiratory distress syndrome (ALI/ARDS) and correlate with
mortality, 736,737 which leads us to hypothesize that inhibition of inflammasome
assembly may be a mechanism for our finding that glibenclamide therapy and
respiratory failure are inversely associated. It is also unsurprising that fewer
genes were found to be differentially expressed in the otherwise healthy diabetic
controls than in the patients with melioidosis, as the inflammasome is only
assembled as part of the acute inflammatory response and so we expect to see
little effect of glibenclamide when there is no active inflammation.
Neutrophil recruitment to an area of tissue inflammation or infection can
also be linked to the inflammasome through the action of IL1β. While IL1β is
not itself a chemoattractant for neutrophils, 738 IL1β up-regulates the expression
of endothelial intercellular adhesion molecules essential for the recruit of
neutrophils to an area of inflammation. 515,516 Neutrophils are a critical
component of the host response to B. pseudomallei in vivo, 326 but may also
drive tissue injury and organ damage. 739–741 The bronchoalveolar fluid of
patients with ALI/ARDS is neutrophil-rich and some animal models of ARDS
are neutrophil-dependent. 739 Glibenclamide has been shown to prevent
neutrophil extravasation and accumulation in animal models, 740,741 and this may
represent indirect evidence that glibenclamide was responsible for reducing the
incidence of respiratory failure in study patients taking this drug. We speculate
that conflicting evidence from previous observational studies of the impact of
diabetes on outcomes from bacterial sepsis may be explained in part by
differences in the management of type 2 diabetes between Europe (metformin
has been available in Britain since 1958 and has traditionally been used first-
line) and the US (where sulfonylureas have often been used first-line 742 and
metformin has only been available since 1995).
114

Glibenclamide has been said to inhibit the inflammasome by blocking
KATP-channels, 743 but this seems unlikely, since the KATP-channel blockers
glipizide and tolbutamide do not have this effect. 744 The two KATP-channels
known to be targeted by the sulphonylureas, ABCC8 (formerly SUR1) and
ABCC9 (formerly SUR2). Neither of these proteins are expressed by
macrophages. 744 The actual target therefore remains to be elucidated. 735,745
The P2X7 receptor has been suggested as a possible target, 746,747 but the
target appears to be downstream of P2X7 735 and there is no sequence or
structural homology between P2X7 and any of the known targets for
glibenclamide. The development of inflammasome-inhibitors that do not have
the hypoglycaemic effects of glibenclamide may depend on unravelling this
pathway. However, enthusiasm for this approach should be tempered by the
finding of one multi-centre trial that direct IL1β blockade does not improve
survival in all-cause sepsis. 363 Our study is also unable to provide guidance on
whether any benefit would be seen were glibenclamide started following the
onset of infection in non-diabetic patients.
Inflammasome function is not restricted to IL1β activation and
inflammasome inhibition is not the only action of glibenclamide. Glibenclamide
is a broad-spectrum inhibitor of ABC-transporters 706–708 and many members of
this family have been implicated in regulating immune function. Glibenclamide
also has a direct effect on neutrophils chemotaxis when administered at a dose
100-times the concentration achievable in humans, 748 although we found no
direct effect of glibenclamide on neutrophils in vitro.
115

4 Glibenclamide reduces interleukin 1β
secretion in a mouse model of
melioidosis
4.1 Introduction
Having found an anti-inflammatory effect of glibenclamide in patients
with melioidosis, we sought to replicate this finding in an experimental model.
The effect we found in the clinic was a general effect on inflammation, but
without one pathway predominating. We therefore hoped than an experimental
model might provide insight into mechanisms underlying the effect of
glibenclamide.
4.1.1 Animal models of melioidosis
The model organism most commonly used to study melioidosis is the
mouse (Mus musculus). The first report of experimental B. pseudomallei
infection in mice was by de Moor et al. in 1932, using a strain described only as
weiss or ‘white’. 242 An important advantage of using a mouse model is the wide
range of tools and reagents available commercially (monoclonal antibodies,
cytokine assays, bioinformatics, and so forth). Indeed, the largest library of gene
knockouts available is in mice, 749,750 and a decision not to use mice means to
forego access to these tools. European legislation requires that the model used
be the least sentient model possible, and there exist established models of both
melioidosis and diabetes in the mouse: this was therefore a consideration in the
selection of a suitable model for this part of the project.
A number of mouse models of melioidosis have been described, but
current published work uses only one of two inbred strains: BALB/c and
C57Bl/6. 751,752 Other strains that have been used include the C3H mouse, 752 the
DBA/2 mouse 752 and the Taylor outbred mouse. 373 Older mouse studies refer to
no specific mouse strain and refer only to colour (white, black or agouti), if at
all.
BALB/c and C57Bl/6 inbred mouse strains form contrasting models of
melioidosis. In 1999, Hoppe et al. surveyed four mouse species (C57Bl/6N,
116

BALB/c, C3H/HeN and DBA/2). 752 He found that BALB/c mice were relatively
susceptible to B. pseudomallei infection, while C57Bl/6N mice were relatively
resistant. Susceptibility of the C3H/HeN and DBA/2 strains were intermediate.
The contrasting susceptibilities of BALB/c and C57Bl/6 has been independently
replicated by other groups. 751,182 The BALB/c mouse is more susceptible to
B. pseudomallei infection, with an LD50 of ~5 cfu. By comparison, the LD50 for
C57Bl/6 mice is consistently two to four orders of magnitude higher, depending
on route of infection. 751,752,182 In C57Bl/6 mice, smaller inocula (e.g., 100 cfu of
B. pseudomallei 576) produce a chronic version of melioidosis. 753
Hoppe found that in BALB/c mice, bacterial loads increased faster in the
first 24 hours than in C57Bl/6 mice. Hoppe also found a higher IgG2a:IgG1 ratio
in C57Bl/6 mice compared to BALB/c mice, which is suggestive of a Th1-type
immune response, 752 because IFNγ promotes IgG2a isotype switching.
This dichotomy between BALB/c and C57Bl/6 has been found in other
infections, e.g., Leishmania major, 754 Mycobacterium tuberculosis, 755 and
Yersinia enterocolitica, 756 where in each instance, BALB/c mice are relatively
susceptible, but C57Bl/6 mice are resistant. * The susceptibility of BALB/c mice
to Leishmania major 526,527 and Y. enterocolitica 756 correlates with a deficiency
in the IFNγ response (traditionally considered a Th1 cytokine). Indeed, it was
initially thought that the contrasting BALB/c and C57Bl/6 models of melioidosis
paralleled the Th1/Th2 paradigm developed for Leishmania. 751
Ullett et al. found that contrary to there being an inadequate IFNγ-
response in BALB/c mice compared to C57Bl/6 mice, IFNγ mRNA levels in liver
and spleen were higher in BALB/c mice than in C57Bl/6 mice. 759 This finding
was confirmed by Koo and Gan using direct measurements of IFNγ as well as
measurements of IFNγ mRNA. 410 Koo and Gan found that splenocytes from
BALB/c mice produced higher concentrations of IL18 and TNFα compared to
splenocytes from C57Bl/6 mice (IL12 concentrations were not different), which
in turn resulted in high concentrations of IFNγ. They hypothesised that the
elevated cytokine response in BALB/c mice is explained by a reduced
responsiveness of phagocytes to IFNγ. This hyporesponsiveness manifested as a
* There exist infections where the converse is true. The C57Bl/6 mouse is more
susceptible to Helicobacter felis 757 and Pseudomonas aeruginosa 758 compared to the
BALB/c mouse.
117

failure of BALB/c phagocytes to kill intracellular bacteria despite adequate IFNγ
stimulation. They felt that the continued presence of intracellular bacteria in
macrophages was responsible for the continued secretion of IL18 and TNFα by
macrophages, which in turn was driving IFNγ secretion. That there is no defect
in the cytokine response was confirmed by Barnes et al., who found that
cytokines downstream of IFNγ, such as CXCL9 (formerly called Mig) and
CXCL10 (formerly called IP-10), were expressed earlier and at higher levels in
BALB/c mice compared to C57Bl/6 mice. 760 Recruitment of neutrophils to sites
of infection in the liver and spleen also occur earlier and in greater numbers in
BALB/c mice compared to C57Bl/6 mice, again consistent with an inability of
BALB/c mice to kill the bacteria, despite apparently adequate cytokine and
cellular responses.
The mechanism underlying the reduced ability of BALB/c mice to kill
intracellular B. pseudomallei is as yet unknown. Miyagi et al. showed that RNS,
and to a lesser extent, ROS were important for intracellular killing in the J774.1
macrophage-like cell line 521 (which is BALB/c derived 761 ). However, subsequent
studies have found that differences in intracellular killing between BALB/c and
C57Bl/6 strains are not explained by differences in RNS and ROS
generation. 519,523
Watanabe et al. have reported that BALB/c macrophages expressed lower
levels of the lysosomal enzyme, β-glucuronidase, in response to macrophage-
activating lipopeptide-2 (a synthetic TLR2 ligand) and to E. coli LPS when
compared to C57Bl/6 macrophages. 762 This is an attractive explanation for the
susceptibility of BALB/c mouse, but the role of lysozymal enzymes has not been
explored in the context of the host response to melioidosis. In humans,
β-glucuronidase deficiency manifests as Sly syndrome, or
mucopolysaccharidosis type VII. The potential association of the BALB/c mouse
with an inherited human disease might prompt caution in the interpretation of
experiments conducted using this strain.
One final comment on the Th1/Th2 balance in melioidosis may be made.
In the context of melioidosis, IFNγ is produced predominantly by NK cells and
CD8 + T lymphocytes, not by CD4 + T helper cells. 360,411 In fact, T helper cells
seem to play little or no role in the initial host response to melioidosis, although
118

they do have a role at later stages in the infection. 411 This picture does not fit
neatly into the conventional Th1/Th2 paradigm, 763 in which T helper cells (Th)
are the primary orchestrators of the cytokine response, even though the
cytokine profile in melioidosis is consistent with the classical Th1 response.
A variety of B. pseudomallei strains are used in animal models, which
hinders the comparison of results from different groups, since different strains
likely have different virulence. Ivo Steinmetz’s group in Greifswald uses the E8
environmental strain of B. pseudomallei collected from Northeast Thailand in
1996 by Andrew Simpson. 32,83,519,752 Despite being an environmental strain, E8
is more virulent than NCTC 7431, which is a clinical isolate. 752 Work published
by Greg Bancroft’s group in London uses the 708a strain, 360,373 which is a
clinical isolate from Thailand. Rick Titball’s group at the Defence Science and
Technology Laboratory (now at Exeter University) uses the 576 764 and
K96243 107 strains. We are not aware of any published studies directly
comparing the virulence of these strains.
The work described here uses the C57Bl/6 model of melioidosis first
described by Wiersinga et al. in 2007. 152,412,765 In this model, 10–14 week-old
animals are inoculated with 500 cfu of B. pseudomallei intranasally. Animals
develop a severe pneumonia, bacteraemia and microabscesses in the liver and
spleen, and median survival is four days. 152 The strain used was B. pseudomallei
1026b, which was originally collected by David Dance in 1993 from a 29-year-
old Thai female rice farmer with bacteraemia, soft tissue, cutaneous lesions,
septic arthritis and splenic abscesses. 93 The strain was donated to Don Woods’
group in Calgary, Canada, in 1996 93 and subsequently transferred from Canada
to Tom van der Poll’s group in Amsterdam, The Netherlands, in 2006, where it
has since been maintained by serial passage in mice.
A variety of other animals have been used to model human melioidosis. In
Whitmore’s 1912 description of the disease, Koch’s postulates were established
using guinea pigs, 8 while Stanton & Fletcher and de Moor et al. both described
rat models of melioidosis. Rats are relatively resistant when inoculated
intraperitoneally (LD50 >10 8 cfu), unless made susceptible by inducing diabetes
with streptozocin (LD50 10 3 –10 4 cfu). 151 Like mice, susceptibility is also
increased by administering the inoculum directly into the respiratory tract (

to B. pseudomallei infection, 35,155 which has led to their use in the study of
bacterial virulence factors 86,88,91 and for the purification of mixed cultures. 140,156
However, some bacterial mutants that are attenuated in mice are not attenuated
in hamsters, 86,88,93,94,767 and it is unclear which model better identifies virulence
factors important to clinical melioidosis.
The first primate model of melioidosis was reported by Miller in 1948
using macaques. 155 More recently, a marmoset model of melioidosis has become
available at the UK Defence Science and Technology Laboratory. 768 An NIH
grant has also been awarded to Don Estes at the University of Texas, Galveston,
to develop a marmoset model of melioidosis (grant no. 5U01AI082103-02 dated
29 Jul 2010). Other models reported in the literature include baboons (Simia
hamadryas) 769 ferrets (Mustela putorius furo), 155 goats (Capra aegagrus
hircus), 770 guinea pigs (Cavia porcellus), 769 pigs (Sus scrofa domestica
Landrace), 771 and roundworms (Caenorhabditis elegans). 772,773
The development of large animal models of melioidosis is critical, because
in the evaluation of a new drug, both the US Food and Drug Administration and
the European Medicines Agency require data from at least two animal models
(small and large) as a pre-condition for the commencement of phase I clinical
trials.
4.1.2 Mouse models of diabetes
Animals have had a prominent role in diabetes research for over a hundred
years. In 1889, Oskar Minkowski removed the pancreas from a dog and found
that it became diabetic, 774 leading the way to Banting and Best’s discovery of
insulin in 1922. 775
We hypothesized that the effect of glibenclamide was independent of its
effect on glucose control. Although the majority of cases of diabetes in Thailand
have type 2 diabetes, we elected to use a model of diabetes in which glucose
concentrations were not responsive to glibenclamide, because alterations to
plasma glucose and insulin concentrations are themselves associated with
changes to the immune response (as reviewed in Chapter 1 of this dissertation).
This review is therefore confined to mouse models of type 1 diabetes.
The best-characterised and most widely-used mouse model of
spontaneous type 1 diabetes is the inbred non-obese diabetic (NOD) mouse,
120

developed in 1974 by Makino at Shionogi Research Laboratories, Japan, from
outbred Swiss mice. 776,777 Most mice of this strain develop some degree of
insulitis (inflammation and destruction of β-islet cells), but progression to
diabetes takes up to 24 weeks and depends on sex and environmental factors.
Under specific pathogen-free (SPF) conditions, 778 80–90% of females are
diabetic by 24 weeks of age; by contrast, only 40% of males are diabetes by 30
weeks of age. The development of diabetes in this strain may be synchronised by
a single dose of intraperitoneal cyclophosphamide 200 mg/kg, 779 which makes
the mice easier to use.
The selection of an appropriate control for the NOD mouse is difficult
because the NOD mouse was developed from an outbred strain, not an inbred
strain, and because the defect is polygenic in origin and poorly defined. 780,781
Both C57Bl/6 mice and BALB/c mice have been used as control strains despite
the fact that neither strain is Swiss-derived. Multiple Swiss-derived control
strains have been developed for the NOD mouse, including the non-obese non-
diabetic (NON) mouse (also developed by Makino), SJL/J, outcrosses of NOD
with diabetes-resistant strains and various NOD congenic strains. The NON
mouse is a particularly troublesome control, because despite its name, it is
prone to obesity and develops impaired glucose tolerance.
Two other mouse models are worth mentioning briefly: the C57Bl/6NJCL-
Insulin2 Akita and the C57BLKS/J-Cpe fat mice.
C57Bl/6NJCL-Insulin2 Akita heterozygous males develop diabetes at
weaning (four weeks of age) and the mouse was described in 1997 as a model for
maturity onset diabetes of the young. 782 Histological examination showed that
the mice had atrophic pancreatic islets deficient in β-cells and the mouse is
therefore a model of type 1 diabetes. The defect is a missense mutation in the
insulin 2 gene (Ins2) on chromosome 7, that renders it non-functional. Note
that Mus musculus possesses two equally-expressed insulin genes: Ins1 is
located on chromosome 19. 783 These mice are insulin-responsive, and non-
diabetic littermates may be used as controls. This mouse model has been
proposed as a replacement for the STZ model. 784
The C57BLKS/J-Cpe fat mouse homozygous males are severely obese by 8
weeks of age, and this is associated with very high circulating insulin levels and
hyperglycaemia, but not severe diabetes. 785 The model was therefore initially
121

described as a model for type 2 diabetes. The defect is a point mutation in
carboxypeptidase E (Cpe) on chromosome 8, which results in abnormal
proinsulin processing. The mouse is therefore insulin-responsive and not
insulin-resistant. The C57BLKS/J background has problems with dental
malocclusion and polycystic kidney lesions which the Jackson Laboratories
solved in 2005 by moving the Cpe fat mutation to a C57Bl/6 background.
A B
Figure 23. A. D-Glucose. B. Streptozocin.
Note.—H = hydrogen; N = nitrogen; O = oxygen. The glucose and streptozocin (STZ)
molecules are oriented to emphasise their structural similarity.
Streptozocin (STZ) was first identified in the late 1950’s by scientists at
Upjohn (U.S. patent 3,027,300 granted 1961) and is a true antibiotic produced
by Streptomyces achromogenes (an actinomycete isolated from soil collected in
Blue Rapids, Kansas). 786
STZ is a glucosamine-nitrosourea compound and, like other nitrosoureas,
it exerts its cytotoxic effects by alkylating DNA. 787 The glucose transporter,
SLC2A2 (previously called GLUT2), is highly expressed by β-cells. The
structural homology between glucose and the glucosamine moiety of STZ
(Figure 23) allows STZ to be preferentially taken up by the β-cell, 788 which then
induces a cytotoxic inflammatory response that destroys the cell. 789 This
explains the specificity of STZ for pancreatic β-cells and STZ therefore has a
clinical application as treatment for β-islet-cell tumours. 790
STZ-treated animals were first described in 1963 by Rakieten et al. 791 as a
model for diabetes and the model is similar to the older alloxan model of
122

diabetes (first described 1943). 792 Both STZ and alloxan are taken up by
SLC2A2, but alloxan kills the cell by producing reactive oxygen species instead
of by DNA alkylation. The main advantage of STZ over alloxan is its (relative)
chemical stability, which makes it easier to use. 788 The main disadvantage of
both chemical models of diabetes is that SLC2A2 is also expressed in the renal
tubule, which means both STZ and alloxan are potentially nephrotoxic. This
nephrotoxicity may be avoided by using multiple smaller doses of STZ (e.g.,
60 mg/kg over five days) instead of fewer large doses (e.g., two doses of
150 mg/kg). 793 In the STZ model of diabetes, glucose levels start to rise a week
after initiating STZ treatment and plateau roughly two weeks after the last dose
of STZ. Only male mice are susceptible to the effect of STZ, which may seem
unusual, but in fact, all mouse models of diabetes exhibit some sexual
dichotomy. The reason for this is not known.
Figure 24. Development of hyperglycaemia in mice treated with STZ.
Sixty mice were injected with STZ 60 mg/kg daily for five days (marked by arrows on
horizontal axis) and compared to 60 controls (treated with citrate buffer pH 4). STZ-
treated mice are shown in red, control mice are shown in black. The horizontal
interrupted line is the 16·7 mM cutoff we used to define diabetes in mice. This number of
mice was chosen to allow for losses due to mice not becoming diabetic or mice dying
from hyperglycaemia prior to inoculation.
An STZ model of melioidosis has previously been described by Woods et
al., who used an infant rat model of melioidosis. 151 In Woods’ report, STZ-
treated infant rats were more susceptible to B. pseudomallei infection than
untreated animals (LD50 ~10 3 –10 4 cfu compared to >10 8 cfu).
123

STZ is stable in solution at pH 4 and rapidly degrades at higher or lower
pH. 791 STZ solutions are therefore made up with cold sodium citrate buffer 50–
100 mM. 793,794 Higher concentrations of sodium citrate are more irritant.
Solutions are not stable for longer than 12 h and must be made up freshly each
day. Woods’ model uses intraperitoneal streptozotocin 80 mg/kg on two
consecutive days, 151 with controls given phosphate buffered saline. It might be
argued that Woods should have used citrate buffer for controls instead of
phosphate-buffered saline in his experiments. Another idiosyncracy of Woods’
report is that the STZ dose he used is not attested elsewhere in the literature: he
references the STZ rat model of Tancrède et al., but Tancrède does not describe
a 80 mg/kg dose. 795
Since starting this project, we have become aware of two other groups
working on STZ-models of melioidosis. The first group is based in Townsville,
Australia, and used a BKS.Cg-Dock7 m+/+ Lepr db /J mouse model of diabetes. 677
This model has a spontaneously arising autosomal recessive mutation in the
leptin receptor (Lepr db ) that arose in the C57BLKS/J (BKS) colony at The
Jackson Laboratories in 1966 and is considered a model of type 2 diabetes. 796
This model is characterized by obesity, hyperglycaemia and insulin insensitivity.
Using this model, Hodgson et al. found that diabetic mice had decreased
survival following subcutaneous B. pseudomallei infection, with 100%
succumbing to the disease in the first four days post-infection, compared to
non-diabetic mice, all of which survived ≥10 days. The group found no defect of
neutrophil or dendritic cell function. The second group is in Kuala Lumpur and
is run by Sheila Nathan at the Universiti Kebangsaan Malaysia; their data is
awaiting publication and was not available at time of writing.
Of the available models for type 1 diabetes, we elected to use the STZ
model of diabetes for practical reasons. The model is easy to use: the phenotype
may be induced on any background, at any time, to produce a cohort of mice
that become diabetic synchronously. At the time the experiments were
designed, we considered the possibility that we might need to use knockout
mice to elucidate mechanisms and the knockouts available to us were all on a
C57Bl/6 background. Furthermore, the murine melioidosis model in
Amsterdam was already established in C57Bl/6 mice.
124

The C57Bl/6 inbred strain separated from the parent C57Bl outbred strain
in 1937 and the colony was moved to the Jackson Laboratories in 1948. The
C57Bl/6J substrain male mouse (Jackson Laboratories) is a model for diet-
induced type 2 diabetes. This is due to a spontaneous 17·8 kb deletion in the
nicotinamide nucleotide transhydrogenase gene, Nnt, that appeared in the
Jackson Laboratories colony some time between 1951 and 1970. 797,798 The
C57Bl/6N substrain (National Institutes of Health, Bethesda, Maryland)
separated from the Jackson Laboratories colony in 1951 and was distributed to
Charles River (Wilmington, Massachusetts) in 1974 (designated C57Bl/6NCrl).
The C57Bl/6N substrain has an intact Nnt gene 799 and is the substrain used for
all mouse experiments described here. The C57Bl/6N substrain is also the
background used by the Knockout Mouse Project. 749,750 This is important,
because there are published studies that have inadvertently used wild-type
C57Bl/6J mice as non-diabetic controls, not realising that this substrain is a
model for diabetes.
4.2 Methods
4.2.1 Mouse experiments
Specified pathogen-free 10-week-old C57Bl/6NCrl mice (Charles River)
were given STZ 60 mg/kg daily for five days. 793 We chose to use a low-dose
regimen because at this dose, STZ does not have a direct toxic effect on the
kidneys, 793 which might otherwise confound our results. STZ 6 mg/ml (Sigma-
Aldrich, Zwijndrecht, The Netherlands) was prepared fresh every day in citrate
buffer (pH 4), then passed through a 0·2 µm polyethersulphone filter (VWR
514-0073) and administered intraperitoneally within 60 minutes of
preparation. Losses were around 20% with this regimen, either from
complications of the hyperglycaemia or because the animals failed to become
diabetic and had to be sacrificed.
Animals were allowed ad libitum access to food (801733 CRM, Tecnilab-
BMI) and water, and were group-housed in polycarbonate cages with 12-hour
light/dark cycles, temperature 18–22°C, humidity 40–65%. Plasma glucose was
checked weekly by cheek puncture 800 (Bayer Contour meter). For 10-week-old
mice not treated with STZ, glucose was 8·75±1·52 mM (mean±standard
125

deviation) and for 16-week-old mice, 9·30±1·64 mM. We defined diabetes as a
plasma glucose ≥16·7 mM. 793
Mice were infected with ~6×10 2 Burkholderia pseudomallei intranasally
4–5 weeks after STZ treatment. In order to mimic the clinical situation seen in
our cohort study, we treated all animals with full-dose antibiotics. Ceftazidime
600 mg/kg (GlaxoSmithKline, Brentford, England) was administered
intraperitoneally starting 24h after inoculation and continued twice daily until
sacrifice. 296
For glibenclamide experiments, glibenclamide 50 mg/kg (Sigma-Aldrich)
in 20% dimethylsulphoxide (DMSO) was administered intraperitoneally
starting 7 days before inoculation and continued until sacrifice; control mice
were given 20% DMSO. This dose of glibenclamide was chosen because it is
equivalent to the highest human dose (20 mg daily to a 50 kg Thai male) after
taking into account differences in pharmacokinetics (human t½ ≈ 8 hours,
mouse t½ ≈ 1 hour).
Eight animals from each group were sacrificed 48, 72 and 96 hours after
inoculation. Mice were sedated with ketamine/medetomidine intraperitoneally
(12·5 mg/20 µg per 100 g weight), then exsanguinated by cardiac puncture.
Death was confirmed by cutting the diaphragm. Glucose measurements were
obtained prior to sedation as medetomidine itself elevates glucose
concentrations.
The lungs were lavaged with three 300 µl aliquots of sterile 0·9% saline
then collected for culture. Lung, liver and spleen were homogenized in four
volumes of sterile saline. Organ homogenates, blood and bronchoalveolar lavage
fluid (BALF) were plated onto blood agar in serial dilutions to quantify bacterial
burdens. The remaining lung homogenate was incubated for 30 minutes with
Greenberger lysis buffer then centrifuged at 650 × g for 10 minutes. The
supernatant was passed through a 0·2 µm filter to remove viable bacteria, then
stored at –20°C for cytokine measurements. Heparinized blood was centrifuged
at 700 × g for 10 minutes; the plasma was filtered and stored at –20°C pending
analysis.
In lung homogenates and BALF, interleukin(IL) 1β, IL6, IL10, CXCL5
(LIX) and tumor necrosis factor-α (TNFα) were measured by sandwich ELISA
(R&D Systems, Oxford, England). Cytokine levels in plasma were measured by
126

particle immunoassay (mouse inflammation cytometric bead array, BD
Biosciences, San Jose, California) because of limited sample quantity. Protein
concentrations were measured using the DC protein assay (Bio-Rad,
Hercules, California).
Statistical analyses were performed on Stata 11 (StataCorp, College Station,
Texas). Bacterial loads and cytokine concentrations were log-normal. Linear
regression was used in preference to ANOVA because it reports not just a p-
value for each group, but also estimates size of difference and permits
interaction testing. Models were fit by the method of variance-weighted least-
squares, because the method does not assume homoscedasticity and because all
explanatory variables could be treated as categorical. Responses were not
assumed to vary linearly with time. Separate p-values were reported for each
time point only when justified by a test for interaction.
Mouse experiments were performed at the Academic Medical Center,
University of Amsterdam and were approved by the animal experimentation
committee there (DIX101725).
4.3 Results
4.3.1 Glibenclamide does not reduce glucose concentrations in the STZ model
of diabetes
The host response is altered by changes in glucose concentration. The
primary action of glibenclamide in diabetes is to reduce blood glucose
concentrations by stimulating insulin secretion by the β-islet cells of the
pancreas. STZ treatment destroys the β-cells of the pancreas, making the
animals hyperglycaemic and unresponsive to the action of glibenclamide. This
allows us to dissect out the separate effects of glibenclamide on glucose and on
the inflammatory response. Blood glucose measurements taken at sacrifice
confirmed that in the STZ model of diabetes, blood glucose concentrations are
not altered by glibenclamide treatment (Figure 25).
127

Figure 25. Glucose concentrations do not fall with glibenclamide
treatment in strepztozocin-treated mice.
The graph shows plasma glucose at sacrifice in mice with STZ-induced diabetes and
compares mice treated with glibenclamide against control mice treated with vehicle
alone. Glibenclamide 50 mg/kg once daily was started 7 days prior to inoculation and
continued until sacrifice; controls were given vehicle alone. All mice were infected with
B. pseudomallei 48, 72 or 96 hours prior to sacrifice.
4.3.2 Glibenclamide does not alter bacterial loads in the lungs
Glibenclamide is structurally related to the sulphonamides, which are
active against B. pseudomallei. If glibenclamide inhibits the growth of
B. pseudomallei, then any differences in the inflammatory response might be
explained by a direct effect on the bacterium and not by an effect on the host
response.
In this model of melioidosis pneumonia, the primary site of infection is the
lungs. We found no effect of glibenclamide on bacterial loads in either lung
tissue or in bronchoalveolar lavage fluid at any of the three time points
measured.
128

Figure 26. Bacterial loads in the lungs and BALF are unaffected by
glibenclamide therapy.
Note.— BALF = bronchoalveolar lavage fluid.
Mice were treated with glibenclamide or vehicle for 7 days prior to inoculation with
650 cfu B. pseudomallei. All mice were treated with ceftazidime 600 mg/kg twice daily
started 24 h after inoculation and continued until sacrifice at 48, 72 or 96 h (n = 8 per
group per time point). Results for glibenclamide-treated animals are indicated in blue,
while results from control animals are indicated in black. Error bars indicate standard
deviations. Bacterial loads were determined by serial dilution on horse blood agar. The
horizontal interrupted line marks the limit of detection for the assay.
4.3.3 Glibenclamide is associated with lower IL1β concentrations
Glibenclamide has previously been reported to inhibit inflammasome
assembly 735 and one of the roles of the NLRP3 inflammasome is the activation
and secretion of interleukin(IL)-1β. We found that in glibenclamide-treated
animals, concentrations of IL1β were 73·3% lower in lung tissue 48 hours after
inoculation compared to controls (p < 0·001, Figure 27).
In lung tissue, IL1β is present in both its mature and immature forms.
Only mature IL1β is secreted into the alveolar space, and IL1β in
bronchoalveolar lavage fluid (BALF) should therefore be a better measure of
inflammasome activation than lung tissue. We found that IL1β concentrations
in BALF were 81·7% lower than controls (p < 0·001) at 48 hours.
129

Figure 27. Cytokine levels in glibenclamide-treated mice
(inflammasome related).
Note.— BALF = bronchoalveolar lavage fluid; IL = interleukin; IFNγ = interferon γ.
Mice were treated with glibenclamide or vehicle for 7 days prior to inoculation with
650 cfu B. pseudomallei. All mice were treated with ceftazidime 600 mg/kg twice daily
started 24 h after inoculation and continued until sacrifice at 48, 72 or 96 h (n = 8 per
group per time point). Results for glibenclamide-treated animals are indicated in blue,
while results from control animals are indicated in black. Error bars indicate standard
deviations. A single p-value is reported for each cytokine unless there is evidence from a
test of interaction that effects at each time point are different. The horizontal interrupted
line marks the limit of detection for the assay.
4.3.4 Glibenclamide is associated with a reduced cellular influx into the lungs
Melioidosis pneumonia is characterised by a florid influx of cells into the
lungs, and these cells are predominantly neutrophils. 765 Although neutrophils
form a critical part of the innate host response to melioidosis, 326 neutrophils
may also contribute to mortality by causing damage to the respiratory tract and
have been implicated in the pathogenesis of the acute respiratory distress
syndrome. 739 IL1β has been implicated in the recruitment of neutrophils to the
inflamed lung, and we hypothesised that glibenclamide might function by
limiting this influx of cells. 515,738
In this study, we found that glibenclamide did not influence the wet weight
of the inflamed lungs or the amount of protein in BALF (both are markers for
capillary leakage) (Figure 28). We did however find that the total number of
130

cells was reduced in glibenclamide-treated animals compared to controls. The
majority of cells in this inflammatory infiltrate were neutrophils, but
macrophages were also present in significant numbers. This difference in cell
counts was most apparent at the 48 h time point (Figure 28).
Figure 28. Lung infiltration in glibenclamide-treated mice.
Note.—BALF = bronchoalveolar lavage fluid. Mice were treated with glibenclamide or
vehicle for 7 days prior to inoculation with 650 cfu B. pseudomallei. All mice were
treated with ceftazidime 600 mg/kg twice daily started 24 h after inoculation and
continued until sacrifice at 48, 72 or 96 h (n = 8 per group per time point). Results for
glibenclamide-treated animals are indicated in blue, while results from control animals
are indicated in black. Error bars indicate standard deviations. A single p-value is
reported for each cytokine unless there is evidence from a test of interaction that effects at
each time point are different.
4.3.5 Glibenclamide is associated with reduced dissemination of
B. pseudomallei from the lungs
Melioidosis is characterised by haematogenous spread from the primary
site of infection to distant organs such as a liver and spleen. 765 B. pseudomallei
is a facultative intracellular organism with an ability to parasitise leukocytes.
We hypothesised that a reduction in leukocyte recruitment to the site of primary
131

infection might therefore hinder the spread of infection to distant organs. We
therefore measured bacterial loads in liver and spleen and found that bacterial
loads were lower at all three time points measured (Figure 29). Consistent with
this, bacterial burdens were also lower in blood, but this is less easy to interpret
statistically, as a large proportion of points were below the limit of detection.
Figure 29. Bacterial loads in blood, liver and spleen are lower in
glibenclamide-treated mice.
Note.—Mice were treated with glibenclamide or vehicle for 7 days prior to inoculation
with 650 cfu B. pseudomallei. All mice were treated with ceftazidime 600 mg/kg twice
daily started 24 h after inoculation and continued until sacrifice at 48, 72 or 96 h (n = 8
per group per time point). Results for glibenclamide-treated animals are indicated in blue,
while results from control animals are indicated in black. P-values reported are for
variance-weight least squares regression. Error bars indicate standard deviations. Bacterial
loads were determined by serial dilution on horse blood agar. The horizontal interrupted
line marks the limit of detection for the assay, which is 3 cfu/50 µl of sample (corrected
for dilution), which is the smallest concentration of bacteria that may be distinguished
from zero with 95% confidence. It was not possible to report a p-value for blood as
>50% of data points were below the limit of detection.
4.3.6 Glibenclamide is not associated with differences in other cytokines
Another role of the NLRP3 inflammasome is the activation and secretion
of IL18. IL18 then stimulates natural killer and CD8 + cells to secrete interferon γ
(IFNγ). Inflammasome inhibition would be expected to inhibit the secretion of
mature IL18, and in turn reduce IFNγ secretion.
Immature IL18 is constitutively expressed at high concentrations in
unstimulated cells. 503 We therefore did not measure IL18 in lung tissue. Instead,
IL18 was measured in BALF, which should represent the mature secreted form
132

of the cytokine only. At the 48-hour timepoint, we found that IL18
concentrations were lower in glibenclamide-treated mice when compared to
controls, however, this difference was not statistically significant (Figure 30).
We found no difference in IFNγ concentrations in lung. IFNγ
concentrations in blood were lower at 48 hours, but this difference was also not
statistically significant (Figure 30). IFNγ was not measured in BALF because
previous experiments have shown it to be at or below the limit of detection in
this model.
Figure 30. IL18 and IFNγ levels in glibenclamide-treated mice.
Note.—BALF = bronchoalveolar lavage fluid; IL = interleukin; IFN = interferon.
Mice were treated with glibenclamide or vehicle for 7 days prior to inoculation with
650 cfu B. pseudomallei. All mice were treated with ceftazidime 600 mg/kg twice daily
started 24 h after inoculation and continued until sacrifice at 48, 72 or 96 h (n = 8 per
group per time point). Results for glibenclamide-treated animals are indicated in blue,
while results from control animals are indicated in black. Error bars indicate standard
deviations. The horizontal interrupted line marks the limit of detection for the assay.
4.3.7 Glibenclamide has no effect on cytokines unrelated to the
inflammasome
In order to determine if the effect of glibenclamide was isolated to IL1β, or
whether it was a general anti-inflammatory effect, we assayed the pro-
inflammatory cytokines, IL6, TNFα and CXCL5. The secretion of these three
cytokines is unrelated to the inflammasome. We also considered the possibility
that glibenclamide might be acting to promote secretion of the anti-
inflammatory cytokine, IL10.
133

Although there were differences in IL6 and TNFα at the 48 h time point,
these differences were small and did not reach statistical significance. Overall,
we found no evidence of an effect of glibenclamide on the secretion of CXCL5,
IL6, IL10 or TNFα (Figure 31).
Figure 31. Levels of cytokine unrelated to the inflammasome in
glibenclamide-treated mice.
Note.—IL = interleukin; TNF = tumour necrosis factor; CXCL5 = chemokine (CXC-motif)
ligand 5. Mice were treated with glibenclamide or vehicle for 7 days prior to inoculation
with 650 cfu B. pseudomallei. All mice were treated with ceftazidime 600 mg/kg twice
daily started 24 h after inoculation and continued until sacrifice at 48, 72 or 96 h (n = 8
per group per time point). Results for glibenclamide-treated animals are indicated in blue,
while results from control animals are indicated in black. Error bars indicate standard
deviations. A single p-value is reported for each cytokine unless there is evidence from a
test of interaction that effects at each time point are different. The horizontal interrupted
line marks the limit of detection for the assay. In blood, concentrations of IL1β, IL10 and
IL12p70 were all below the limit of detection (data not shown).
134

4.4 Discussion
We have presented evidence for an anti-inflammatory effect of
glibenclamide in a diabetic mouse model of melioidosis and this anti-
inflammatory effect appears specific to IL1β. This is consistent with the results
of previous in vitro and in vivo reports. 735,801,802 To the best of our knowledge,
this is the first time an inhibitory effect of glibenclamide on IL1β secretion has
been described in any sepsis model.
A limitation of clinical studies of infection has always been that the
interplay between the burden of infection and the inflammatory response has
been difficult to tease out. Cytokine levels may be higher because bacterial
burdens are higher, or bacterial burdens may be higher because of an ineffective
innate host response. 376 Animal studies allow us to control the size of the initial
inoculum and therefore the initial burden of infection. In this study, we found
that bacterial loads in the lungs (the primary site of infection in our model) were
not different between glibenclamide-treated and untreated mice. This excludes
the possibility that the differences in IL1β levels in the lungs are due to an effect
of glibenclamide on the burden of infection in the lungs.
We found that glibenclamide limits the cellular influx into the lungs, with
effects on both neutrophil and monocyte numbers. Neutrophils are a crucial
part of the host response to melioidosis, 326 but neutrophils have also been
implicated in the pathogenesis of the acute respiratory distress syndrome. 739
Glibenclamide was not associated with a complete ablation of the neutrophil
response (neutrophils are not a normal constituent of BALF taken from the
healthy lung). Glibenclamide was only associated with a reduction in the
magnitude of the neutrophil response and not its complete ablation. This
balance between ‘enough’ and ‘too much’ is a common theme in the
pathogenesis of sepsis and its complications. 803 A reduction in neutrophil influx
into the lungs is consistent with our finding that glibenclamide is associated
with a lower incidence of respiratory failure in patients with melioidosis
(Chapter 2).
It is also worth noting that B. pseudomallei is a facultative intracellular
pathogen and that it is capable of parasitizing macrophages. 79–81 Bacterial
135

urdens in blood, liver and spleen were lower in our model, suggesting that
glibenclamide reduces bacterial dissemination. The role of neutrophils and
macrophages in the dissemination of B. pseudomallei is not known, but it is
possible to speculate that the reduced influx of leukocytes into the lungs reduces
the number of cells parasitized by the bacterium and therefore impairs the
ability of B. pseudomallei to disseminate.
We found no effect on cytokines unrelated to the inflammasome.
Specifically, there was no effect on CXCL5, IL6, IL10 or TNFα secretion, which
is again consistent with previous reports in the literature 735,804 and rules out a
more general anti-inflammatory effect. In this model, all animals were treated
with full-dose ceftazidime in order to simulate the clinical study that motivated
this experiment. 329 Under this regimen, all animals showed signs of recovery by
96 hours and the fact that no differences were seen in cytokine responses at
later time points is therefore not surprising.
The mechanism by which glibenclamide is able to inhibit IL1β secretion is
not known. Lamkanfi et al. demonstrated that glibenclamide is able to inhibit
assembly of the NLRP3 inflammasome in response to stimulation with
lipopolysaccharide (LPS) and adenoside triphosphate (ATP). However, when
S. Typhimurium was used as a stimulus, glibenclamide was not able to prevent
inflammasome assembly, as measured by caspase 1 cleavage (caspase 1 being a
critical component of the NLRP3 inflammasome). This suggests that there are
alternative pathways which permit the assembly of the NLRP3 inflammasome,
and the specific pathway that glibenclamide blocks is bypassed when
macrophages are confronted by a complex stimulus such as a whole bacterium.
In our model, glibenclamide-treatment reduced IL1β secretion in BALF,
but did not block it completely. This means that there must have been some
inflammasome assembly despite glibenclamide treatment, because only mature
IL1β is secreted into BALF. Immature IL18 is also cleaved by the
inflammasome, so our finding that inflammasome assembly is not completely
blocked by glibenclamide in melioidosis is supported by the fact that we
observed only a modest reduction in IL18 that was not statistically significant.
Our study does not exclude other mechanisms for the action of
glibenclamide. Aside from an effect on the inflammasome, Hamon et al. have
presented evidence that glibenclamide may block the secretion of mature
136

IL1β. 804 Glibenclamide has also been shown to have other anti-inflammatory
effects, including the prevention of ischaemic reperfusion injury, 740,741 and an
enhancement of intracellular killing of Leishmania parasites. 805
In conclusion, we have replicated our clinical finding that glibenclamide is
associated with reduced inflammation in melioidosis. Specifically, we found
reductions in IL1β secretion, cellular infiltration into the lungs and
dissemination to distant organs. Glibenclamide has previously been shown to
prevent IL1β maturation and secretion by inhibiting inflammasome assembly,
but the presence of detectable IL1β in BALF coupled with smaller-than-expected
effects on IL18 and IFNγ means inflammasome inhibition is incomplete, or that
this may not be the mechanism by which glibenclamide acts to reduce IL1β in
the context of melioidosis.
137

5 Final discussion and conclusion
We have shown that patients with a pre-existing diagnosis of diabetes have
a lower mortality than patients without diabetes and patients with
hyperglycaemia or undiagnosed diabetes. There is no protective effect of
diabetes per se: instead, the mortality benefit seen is associated specifically with
glibenclamide therapy. There was no evidence for an effect of diabetes on
mortality after adjusting for the effect of glibenclamide. We also found that
glibenclamide appears to protect from respiratory failure.
In a gene expression study of total peripheral blood leukocytes in
melioidosis patients, glibenclamide was associated with a broad decrease in
inflammation, with no one pathway predominating. We therefore used a mouse
model to dissect out the effect of glibenclamide on the host response and found
that glibenclamide reduces IL1β secretion in melioidosis with no effects on other
cytokines. The reduction in IL1β secretion is also associated with a reduced
cellular influx into the lungs, which may help explain the clinical association
with a reduced incidence of respiratory failure, because neutrophils have been
implicated in the pathogenesis of the acute respiratory distress syndrome. There
was also an association with decreased bacterial dissemination to distant sites,
which we speculate is related to the reduced cellular influx.
The mechanism by which glibenclamide reduces IL1β production by
monocytes is not known. In the context of LPS stimulation, glibenclamide has
been shown to block inflammasome assembly in isolated macrophages,
however, we have only equivocal evidence for an effect of glibenclamide on the
inflammasome in the context of melioidosis.
It may be argued that a knowledge of mechanism is not required for a drug
to enter therapeutic use. The primary indication for glibenclamide is the
treatment of diabetes and glibenclamide stimulates insulin production
regardless of plasma glucose concentrations. The administration of
glibenclamide to the critically ill but euglycaemic patient may therefore induce
potentially fatal hypoglycaemia. While hypoglycaemia is managed with relative
ease in the setting of the Western intensive care unit, the use of glibenclamide is
138

dangerous in a resource poor setting where close monitoring is not so readily
available.
5.1 Cardiac effects of glibenclamide
A further problem with the use of glibenclamide is its effect on the heart.
In 1970, the University Group Diabetes Program (UGDP) reported that
cardiovascular mortality was 2½-times higher in patients treated with the
sulphonylureas (principally, tolbutamide) compared to patients treated with
placebo and that limb of the trial was terminated early. 2 These findings created
controversy for a number of reasons: first, because they were unexpected and,
second, because the excess mortality was seen at only three centres. These
findings were not unchallenged 806,807 and the criticism surrounding their report
was heated. The subsequent debate was governed in part by emotion, 808 but also
by commercial interests centred around tolbutamide (Orinase®, then
manufactured by Upjohn). Subsequent studies have reported mixed
conclusions, with some supporting the effect of sulphonylureas on
cardiovascular mortality and some not. The interested reader is referred to the
fully referenced record of that debate which has been published elsewhere. 808
The response to the UGDP results was mixed. Phenformin (the only
biguanide then available in the US, and also the only oral alternative at that
time) was withdrawn in the mid-1970’s, on the back of evidence that it caused
fatal lactic acidosis (also from the UGDP). The lack of oral alternatives and
mixed evidence from subsequent trials meant a partial return of confidence in
the use of sulphonylureas. 809 Doubts have lingered and the FDA mandated
package insert for all sulphonylureas sold in the US still contains a bold-print
warning about cardiovascular risks. 742
Follow up studies looking at the effect of sulphonylureas of cardiovascular
mortality have found that it is not a class effect, and that in addition to
tolbutamide, glibenclamide is a prime offender. 810–812 Other members of the
class, such as gliclazide do not appear to have this problem.
Glibenclamide is thought to increase the risk of cardiovascular disease by
blocking ischaemic pre-conditioning. 813 Ischaemic pre-conditioning is a
phenomenon in heart muscle where brief episodes of ischaemia render the heart
more resistant to subsequent ischaemia. Diabetics often have diffuse coronary
139

disease which means diabetic patients who have a heart attack are likely to have
had smaller ischaemic insults that presage the heart attack and later partially
protect it from the effect of the larger, more serious event through ischaemic
pre-conditioning. Ablation of the ischaemic pre-conditioning response removes
this acquired resistance to ischaemia. There is indeed evidence from the warm-
up phenomenon (patients with coronary artery disease are given two
consecutive exercise tests, the relevant outcomes being symptomatic angina and
electrocardiographic changes on the second test) 814,815 and from human
angioplasty studies that glibenclamide is able to block ischaemic pre-
conditioning. 816–818 The experimental evidence therefore supports the
observational evidence that glibenclamide predisposes to cardiovascular
mortality.
The effect on ischaemic preconditioning is not seen across the whole
family of sulphonylureas. The second-generation sulphonylurea, glipizide, did
not block ischaemic preconditioning in rabbits, 819 and the meglitinide analogues
(“glinides”) were originally marketed on the basis that they have a lower affinity
for cardiac receptors (however, the post-marketing evidence for that is mixed:
glimepiride appears not to block ischaemic pre-conditioning, 817,818,820 but
repaglinide does 821 ).
There is a second mechanism by which sulphonylureas might increase
cardiovacular mortality, which is through their effects on smooth muscle. The
second-generation sulphonylureas, glibenclamide and gliclazide, and (to a lesser
extent) the glinide, glimepiride, all reduce coronary artery blood-flow, increase
coronary resistance, depress the mechanical activity of the heart and increase
myocardial oxygen extraction. 822 It may be noted that this vasopressor effect is
precisely that which made glibenclamide an attractive candidate for the
adjunctive treatment of sepsis in the 1990’s.
These data should give us pause when considering the possibility of using
glibenclamide to prevent infection-related mortality in patients with diabetes,
because any gain from preventing infection-related mortality might be wiped
out by an increase in cardiovascular mortality.
140

5.2 Separating the anti-inflammatory effects of glibenclamide
from its effects on insulin secretion
Purely chemical considerations mean that separating the anti-
inflammatory effect of glibenclamide from its hypoglycaemic effect may not be
possible. To explain why this is so requires elaborating on the chemical
structure of glibenclamide, its metabolites, and the pharmacological properties
of the related compound, meglitinide.
The two major metabolites of glibenclamide are 4-trans-hydrocyclohexyl
glibenclamide (M1) and 3-cis-hydrocyclohexyl glibenclamide (M2 or M2b) 823
and account for 36% and 15% of the total, respectively (Figure 32). Experiments
in rabbits suggested that these two compounds are biologically inactive (they
are 1/400th and 1/40th as active, respectively, and data from rats show that 4-
trans-hydrocyclohexyl glibenclamide is only 1/6·5 as potent as
glibenclamide). 824 However, Rydberg et al. reported that in healthy human
volunteers, glibenclamide, M1 or M2, all have equivalent biological activity. 823
Indeed, glibenclamide is much more likely to cause hypoglycaemia than its half-
life would suggest (particularly in patients aged 70 years and older), 825
presumably because of its active metabolites.
Glibenclamide metabolites are excreted 50% in urine and 50% in bile. This
is qualitatively different from that of other sulphonylureas, which are excreted
primarily in the urine. Other metabolites that have been described are 4-cis-
(M2a), 3-trans- (M3), 2-trans-hydrocyclohexylglibenclamide (M4), and ethyl-
hydroxyglibenclamide (M5) (Figure 32). 826 Although the procedure for
synthesising these metabolites has been described, 827 these compounds are not
available commercially and their pharmacological and pharmacokinetic
properties have not been studied clinically.
141

General
structure:
Glibenclamide: R1 =
M1: R1 =
M2a: R1 =
M2b: R1 =
M3: R1 =
M4: R1 =
M5: R1 =
142
R2 = –H
R2 = –H
R2 = –H
R2 = –H
R2 = –H
R2 = –H
R2 = –OH
Figure 32. Structure of glibenclamide and its metabolites.

Figure 33. Structure of meglitinide.
The molecule is oriented to emphasise its homology with glibenclamide (above). IUPAC
name: 4-[2-[(5-chloro-2-methoxy–benzoyl)amino]ethyl]benzoic acid (molecular weight,
333·8 g/mol). At present, meglitinide is not available commercially or clinically, but the
compound was previously available as a gift from Hoechst (now part of Sanofi-Aventis).
The patent on meglitinide is still in force, so it is not possible to study this compound
without the cooperation of Sanofi-Aventis.
Nevertheless, certain properties may be deduced from the chemical
structure of these compounds. Lamkanfi et al. have found that inflammasome
inhibition by glibenclamide is not dependent on the cyclohexyl ring of
glibenclamide. 735 Metabolites M1, M2a, M2b, M3 and M4 (M1–4) only differ
from glibenclamide in that the cyclohexyl ring has been hydroxylated. If
Lamkani is right, then the five compounds, M1–4, would all be predicted to
inhibit inflammasome assembly. M5 differs from glibenclamide in that one of
the carbons between the two aromatic rings patients has been hydroxylated and
the effect of this modification is unknown.
Meglitinide is an insulin secretagogue that is not a sulphonylurea, but like
the sulphonylureas, meglitinide will stimulate pancreatic β-islet cells to secrete
insulin (Figure 33). 828 Comparison of the meglitinide with M1–4 reveals that the
half of the glibenclamide molecule containing the two aromatic rings is almost
identical to meglitinide. This means that metabolites M1–4 should all possess
insulin secretagogue properties also, because the meglitinide moiety is
unaltered. Evidence that this is true in humans is provided by Rydberg’s clinical
study, cited above. 823 This means that the anti-inflammatory part of the
glibenclamide molecule is probably also responsible for at least part of its
hypoglycaemic effect.
143

5.3 Suggestions for future research
The unwanted effects of glibenclamide (hypoglycaemia and adverse
cardiovascular effects) mean that glibenclamide is an unattractive candidate
drug for the prevention of melioidosis-related mortality. Further advances will
only occur when the precise mechanism underlying the action of glibenclamide
has been elucidated, because only then may rational therapies be designed.
IL1β is a product of the macrophage inflammasome and the
inflammasome has already been studied in bone marrow-derived macrophage
models of B. pseudomallei infection. 473,474 Future studies using these systems
may help either to confirm or refute the role of the inflammasome in the effect
of glibenclamide on the host response to melioidosis.
144

A Appendix. Microarray results
Appendix Table 1. List of genes differentially expressed in patients with melioidosis on glibenclamide at the time of
admission
ProbeID Symbol Gene RefSeq ID Absolute Corrected Comment
fold change p-value
Immune-related
5860377 ABCG1 ATP-binding cassette protein G1 NM_016818.2 +1·55 0·03 Regulation of Toll-like receptor expression
5420440 ABP1 Amiloride-sensitive amine oxidase NM_001091.2 –2·80 0·07 Postulated to be involved in endocytosis
6220195 BATF Basic leucine zipper transcription factor, NM_006399.2 –1·54 0·05 Transcription factor elevated in B-cells following EBV
ATF-like
infection
3140487 BMX BMX non-receptor tyrosine kinase NM_001721.4 –1·60 0·07 Regulates TLR-4-mediated IL-6 production.
6270681 BST1 Bone marrow stromal cell antigen 1; CD157 NM_004334.1 –1·56 0·05 ?Regulation of B-cell maturation
1070367 C19orf59 or Mast cell expressed membrane protein 1; NM_174918.2 –2·95 0·04 Hypothesized to be involved in mast cell function or
MCEMP1 chromosome 19 open reading frame 59
differentiation
840068 C3AR1 C3a anaphylatoxin chemotactic receptor NM_004054.2 –2·30 0·03 Inflammatory response and cell motility
2000128 C4BPA Complement component 4 binding protein,
α subunit
NM_000715.3 –2·42 0·03 Classical pathway of complement cascade
6590682 CCL3 or Chemokine (C-C motif) ligand 3; or
NM_002983.1 –1·67 0·03 Neutrophil recruitment and activation
MIP1A
macrophage inflammatory protein 1A
1300465 CD177 Neutrophil-specific antigen 1 NM_020406.2 –2·42 0·07 Neutrophil development and differentiation
1240450 CD27 CD27 NM_001242.4 +1·62 0·04 Generation and long-term T-cell immunity. Regulates
B-cell activation and immunoglobulin synthesis
4610110 CD300LB CD300 molecule-like family member B,
TREM5
NM_174892.2 +1·57 0·04 Activating receptor expressed on myeloid cells
4290097 CD99 CD99 protein NM_002414.3 +1·61 0·03 Cell surface glycoprotein: leukocyte migration, T-cell
145
adhesion & apoptosis
5700753 CEACAM1 Carcinoembryonic antigen-related cell NM_00102491 –2·03 0·05 Myeloid cell adhesion to endothelium and
adhesion molecule 1 (=biliary glycoprotein) 2.1
angiogenesis
1260228 CLEC5A C-type lectin domain family 5, member A NM_013252.2 –1·85 0·06 Myeloid cell activation
4260484 COMMD6 COMM domain containing 6 NM_203497.2 –1·58 0·04 Inhibits NFκB

1050025 CR1 Complement receptor 1 NM_000573.3 –1·67 0·04 C3b and C4b binding
6450543 CST3 Cystatin C NM_000099.2 +1·84 0·03 Inhibits lysosomal cysteine proteases
4150189 CTSL1 Cathepsin L1 NM_001912.3 –1·73 0·04 Controls neutrophil elastase activity
6270553 CXCL10 Chemokine (C-X-C motif) ligand 10 NM_001565.2 –2·08 0·08 Monocyte, T-cell and NK cell activation
2340220 CD55 Complement decay accelerating factor NM_000574.2 –1·52 0·07 Complement regulation
3360279 CFH Complement factor H NM_00101497
5.1
–1·55 0·07 Complement cascade regulation
2810767 EBI3 Epstein-Barr virus induced 3 NM_005755.2 –2·33 0·05 Positive regulation of IFN-γ response
5560471 FCAR Receptor for Fc fragment of IgA NM_133280.1 –1·58 0·04 IgA receptor
6620209 FCGR1B or Receptor for Fc fragment of IgG, high affinity NM_00100434 –1·81 0·08 IgG receptor
CD64
Ib
0.1
2350066 HLA-A Human leukocyte antigen-A NM_002116.5 +1·61 0·03 MHC class I molecule
4250364 HPGD 15-hydroxy prostaglandin dehydrogenase NM_000860.3 –1·53 0·06 Prostaglandin metabolism
6040647 HSH2D Hematopoietic SH2 domain containing NM_032855.2 +1·61 0·03 Target of T-cell receptor and CD28 activation.
Function unknown.
2000148 IFIT1 Interferon-induced protein with
NM_001548.3 +2·48 0·08 Interferon-mediated innate immune response
tetratricopeptide repeats 1
4730747 IGLL1 Immunoglobulin lambda-like polypeptide 1 NM_020070.2 +1·74 0·04 Pre-B cell receptor, controls B-cell proliferation and
differentiation
3370326 IL18R1 Interleukin-18 receptor 1 NM_003855.2 –2·17 0·08 IFN-γ mediated immune responses
1340743 IL8 Interleukin-8 NM_000584.2 –2·10 0·03 Neutrophil chemotaxis
2360719 IRAK3 Interleukin-1 receptor-associated kinase 3 NM_007199.1 –2·13 0·04 Monocyte tolerance to endotoxin
2490411 ITGB5 Integrin β5 NM_002213.3 +1·52 0·04 Cell adhesion
4150184 LAIR1 or Leukocyte-associated immunoglobulin-like NM_021706.2 –1·78 0·04 Inhibits NK cells, T-cells and B-cells
CD305 receptor 1
6110037 LILRA3 Leukocyte immunoglobulin-like receptor,
subfamily A (without TM domain), member 3
NM_006865.2 –1·65 0·04 ?Inhibits monocyte and B-cell function
5890095 LILRA5 Leukocyte immunoglobulin-like receptor, NM_021250.2 –1·59 0·06 Monocyte activation
subfamily A (with TM domain), member 5
6420450 LIME1 Lck-interacting transmembrane adaptor 1 NM_017806.1 +1·55 0·09 Expressed on T-cell membranes, and downmodulated
on activation
6840408 LY86 Lymphocyte antigen 86 NM_004271.3 +1·55 0·07 Humoral immune response
70167 LY96 Lymphocyte antigen 96 NM_015364.2 –1·92 0·03 Associated with TLR4
5130767 MAPK14 Mitogen-activated protein kinase 14 NM_001315.1 –1·52 0·08 Activated by pro-inflammatory cytokines
7550066 MERTK c-mer proto-oncogene tyrosine kinase NM_006343.2 –1·58 0·08 Mediates macrophage phagocytosis of apoptotic cells
5670674 MME Membrane metallo-endopeptidase NM_000902.3 +3·33 0·03 Lymphocyte activation; Inactivates peptide hormones
146

1240228 PAG1 Phosphoprotein associated with
glycosphingolipid microdomains 1
NM_018440.3 –1·53 0·03 ?regulation of T-cell activation
2230563 PPBP Platelet basic protein NM_002704.2 +1·68 0·05 CXC chemokine. Neutrophil chemoattractant and
activator.
7040470 PSMA6 Proteasome (prosome, macropain) subunit,
α type, 6
NM_002791.1 –1·52 0·04 Processing of MHA class I peptides
4890707 PVRL2 or Poliovirus receptor-related 2 (herpesvirus NM_002856.2 –1·56 0·04 Enhances NK cell-mediated lysis
CD112 entry mediator B)
130681 RNASE6 Ribonuclease, RNase A family, k6 NM_005615.4 +1·56 0·05 Expressed in neutrophils and monocytes
6220438 RPS6KA5 Ribosomal protein S6 kinase, 90kDa,
polypeptide 5
NM_004755.2 +1·65 0·04 MAP kinase signaling pathway
2490195 S100A10 S100 calcium binding protein A10 NM_002966.1 +1·66 0·03 Cell cycle progression and differentiation
1410221 S100A12 S100 calcium binding protein A12 NM_005621.1 –2·62 0·05 Myeloid-related protein
1510424 S100P S100 calcium binding protein P NM_005980.2 –2·27 0·09 RAGE ligand
1410181 SAMSN1 SAM domain, SH3 domain and nuclear
localization signals 1
NM_022136.3 –1·97 0·07 B-cell activation
6040577 SLAMF8 SLAM family member 8 NM_020125.2 –1·50 0·07 Lymphocyte activation
3120543 SLC39A8 Solute carrier family 39 (zinc transporter),
member 8
NM_022154.5 –1·69 0·06 Cellular zinc import at onset of inflammation
1990300 SOCS1 Suppressor of cytokine signaling 1 NM_003745.1 –1·67 0·07 Regulation of IFNγ response
160019 SORT1 Sortilin 1 NM_002959.4 –1·62 0·07 Trans-Golgi network transmembrane protein, antiapoptotic
and B-cell survival
3390612 TLR8 Toll-like receptor 8 NM_016610.2 –1·60 0·03 Pathogen recognition
2640301 TNF Tumor necrosis factor (TNF superfamily,
member 2)
NM_000594.2 –1·53 0·03 Proinflammatory cytokine
6350632 TSC22D3 TSC22 domain family, member 3 NM_198057.2 +1·53 0·03 Stimulated by glucocorticoids and IL-10. Antiinflammatory
5910113 VCAN Versican core protein NM_004385.2 +1·62 0·03 Leukocyte aggregation and pro-coagulant
5310754 VNN1 Vanin 1 NM_004666.1 –2·19 0·07 Hematopoeitic cell trafficking
7100161 VNN2 Vanin 2 NM_004665.2 –1·54 0·04 Transendothelial migration, monocyte activation
Coagulation
2230241 F13A1 Coagulation factor XIII, subunit A NM_000129.3 +1·89 0·04 Clotting cascade
2340220 CD55 Complement decay accelerating factor NM_000574.2 –1·52 0·07 Complement regulation
3360279 CFH Complement factor H NM_00101497
5.1
–1·55 0·07 Complement cascade regulation
147

Carbohydrate metabolism
4810438 MAN2B2 Mannosidase α, class 2B, member 2 NM_015274.1 +1·54 0·006 Carbohydrate metabolism
290348 POFUT1 GDP-fucose protein O-fucosyltransferase 1 NM_015352.1 +1·55 0·03 Carbohydrate metabolism
3390129 C9orf103 Chromosome 9 open reading frame 103 NM_00100155 –1·67 0·05 Probable glucokinase
940274 GK Glycerol kinase NM_000167.3 –1·62 0·05 Glycerol phosphorylation
670528 GYG1 Glycogenin 1 NM_004130.2 –1·62 0·08 Regulates glycogen synthesis
3060612 HK3 Hexokinase 3 (white cell) NM_002115.1 –1·93 0·04 Glucose phosphorylation
6620689 MTHFD2 Methylenetetrahydrofolate dehydrogenase
(NADP+ dependent) 2,
methenyltetrahydrofolate cyclohydrolase
7040181 PFKFB2 6-phosphofructo-2-kinase/fructose-2,6biphosphatase
2
1470601 PFKFB3 6-phosphofructo-2-kinase/fructose-2,6biphosphatase
3
1.1
NM_00104040
9.1
148
–1·52 0·04 Mitochondrial bifunctional enzyme
NM_00101805
3.1
–2·30 0·05 Glycolysis/gluconeogenesis
NM_004566.2 –1·76 0·04 Glycolysis/gluconeogenesis
4830504 PGM2 Phosphoglucomutase 2 NM_018290.2 –1·52 0·04 Glucose and glycogen metabolism
6180465 SLC2A6 Solute carrier family 2 (facilitated glucose
transporter), member 6
NM_017585.2 –1·77 0·03 Hexose transport across cell membrane
3420332 SLC37A3 Solute carrier family 37 (glycerol-3phosphate
transporter), member 3
NM_032295.2 –1·50 0·05 Transmembrane sugar transport
Oxidative stress and DNA damage
130181 ANKRD22 Ankyrin repeat domain 22 NM_144590.2 –2·93 0·06 Mediates protein-protein interactions
4880673 GADD45A Growth arrest and DNA-damage-inducible,
alpha
NM_001924.2 –2·02 0·05 Stress response
2600382 GPX4 Glutathione peroxidase 4 NM_00103984
7.1
+1·57 0·02 Reduces peroxide
6400392 GSTM1 Glutathione S-transferase µ1 NM_000561.2 +2·35 0·03 Deals with oxidative stress
6130168 GSTM2 Glutathione S-transferase µ2 NM_000848.2 +2·38 0·03 Deals with oxidative stress
150400 LBA1 Lupus brain antigen 1 NM_014831.1 +1·52 0·04 DNA repair
5050537 OPLAH 5-oxoprolinase (ATP-hydrolyzing) NM_017570.2 –1·75 0·04 Glutathione metabolism
2510551 STK3 Serine/threonine kinase 3 (STE20 homolog,
yeast)
NM_006281.2 –1·60 0·03 Stress response

DNA replication, cell cycle regulation, cell differentiation and proliferation
940288 BAZ1A Bromodomain adjacent to zinc finger
domain, 1A
NM_182648.1 –1·60 0·03 DNA replication
7400279 CCNA1 Cyclin-A1 NM_003914.2 –1·84 0·05 Cell cycle regulation
6220674 CCPG1 Cell cycle progression 1 NM_020739.2 –1·54 0·03 Cell cycle regulation
1780564 CDC42 Cell division cycle 42 NM_00103980
2.1
–1·59 0·03 Regulation of cell cycle progression
6200064 CECR1 Cat eye syndrome chromosome region,
candidate 1
NM_177405.1 +1·79 0·06 Adenosine deaminase; putative growth factor
4490528 CKAP4 Cytoskeleton-associated protein 4 NM_006825.2 –1·79 0·05 Cell cycle regulation
6510554 DACH1 Dachshund homolog 1 (Drosophila) NM_080760.3 –1·76 0·03 Cell fate in eye development
7560288 DCUN1D3 DCN1, defective in cullin neddylation 1,
domain containing 3 (S. cerevisiae)
NM_173475.1 –1·51 0·03 Cell cycle regulation
4220605 ETS2 v-ets erythroblastosis virus E26 oncogene
homolog 2 (avian)
NM_005239.4 –1·68 0·07 Stem cell development
610164 FYN FYN oncogene NM_153047.1 +1·50 0·05 Tyrosine kinase oncogene
6180427 G0S2 G0/G1switch 2 NM_015714.2 –1·53 0·06 Cell cycle regulation
6130136 GPR177 G-protein coupled receptor 177 NM_00100229
2.1
+1·57 0·04 Required for WNT3A expression (cell fate regulation)
5490768 GPR56 G protein-coupled receptor 56 NM_201525.1 +1·61 0·10 Cell fate regulation and neurogenesis
4120750 OBFC1 Oligonucleotide/oligosaccharide-binding
fold containing 1
NM_024928.3 –1·56 0·04 Increases DNA-polymerase-template affinity
4260253 OBFC2A Oligonucleotide/oligosaccharide-binding NM_00103171 –1·51 0·03 ?DNA metabolism
fold containing 2A
6.1
2470520 RBMS1 RNA binding motif, single stranded
interacting protein 1
NM_002897.3 –1·83 0·03 DNA/RNA processing
6960022 SEPT5 Septin 5 NM_002688.4 +1·58 0·08 Cell cycle regulation
3830735 UPB1 Ureidopropionase β NM_016327.2 –1·59 0·07 Pyrimidine degradation
7570673 UPP1 Uridine phosphorylase 1 NM_003364.2 –1·61 0·08 Uridine phosphorylation
149

Appendix Table 2. List of genes differentially expressed in glibenclamide-treated diabetic patients who are
otherwise healthy
ProbeID Symbol Gene RefSeq ID Absolute
fold
change
150
Corrected
p-value
Comment
Inflammation-related
2030142 CLC Charco-Leyden crystal protein NM_001828.4 +2·01 0·05 Lysophospholipase expressed by eosinophils and
basophils
2260731 ERAP2 Endoplasmic reticulum aminopeptidase 2 NM_022350.2 +1·58 0·05 Trimming peptides for MHC class I presentation
940519 GPR44 G protein-coupled receptor 44 NM_004778.2 +1·57 0·05 Chemoattractant receptor expressed on Th2 cells
510079 HLA-DRB4 HLA-DRB1 MHC class II antigen NM_021983.4 –1·58 0·05 Antigen presentation
620544 HLA-DRB6 HLA-DRB6 MHC class II antigen NR_001298.1 –2·50 0·05 Antigen presentation
4210458 KIR2DL1 Natural killer cell immunoglobulin-like
receptor, 2 domains, long cytoplasmic chain, 1
NM_014218.2 –1·54 0·06 NK cell modulation
Neutrophil function
1500735 CTSG Cathepsin G NM_001911.2 –1·84 0·05 Neutrophil azurophil granule protease
4540239 DEFA1 Neutrophil defensin 1 NM_004084.2 –2·49 0·05 Neutrophil microbicidal peptide
2970747 DEFA3 Neutrophil defensin 3 NM_005217.2 –2·49 0·06 Neutrophil microbicidal peptide
6550164 DEFA4 Neutrophil defensin 4 NM_001925.1 –2·20 0·05 Neutrophil microbicidal peptide
7650497 ELA2 Neutrophil elastase 2 NM_001972.2 –1·93 0·05 Neutrophil azurophil granule protease
7150170 LOC728358 Neutrophil defensin 1 precursor NM_001042500.1 –2·58 0·05 Neutrophil microbicidal peptide
2470364 PRSS33 Serine protease 33 NM_152891.2 +1·85 0·05 Serine protease expressed in neutrophils and
macrophages
Note.—All genes in this table have a fold change in expression greater than 1·5 and a corrected p-value less than 0·10 (Benjamini-Hochberg);
when more than one probe for the same gene was positive, only the probe with the most significant corrected p-value is displayed. Only ontologies
with a significance level

References
1. The effect of intensive treatment of diabetes on the development and
progression of long-term complications in insulin-dependent diabetes
mellitus. The Diabetes Control and Complications Trial Research Group. N.
Engl. J. Med 329, 977-986 (1993).
2. Meinert, C.L., Knatterud, G.L., Prout, T.E. & Klimt, C.R. A study of the
effects of hypoglycemic agents on vascular complications in patients with
adult-onset diabetes. II. Mortality results. Diabetes 19, Suppl:789-830
(1970).
3. Knatterud, G.L., Meinert, C.L., Klimt, C.R., Osborne, R.K. & Martin, D.B.
Effects of hypoglycemic agents on vascular complications in patients with
adult-onset diabetes. IV. A preliminary report on phenoformin results.
JAMA 217, 777-784 (1971).
4. Intensive blood-glucose control with sulphonylureas or insulin compared
with conventional treatment and risk of complications in patients with type
2 diabetes (UKPDS 33). UK Prospective Diabetes Study (UKPDS) Group.
Lancet 352, 837-853 (1998).
5. Peacock, S.J. Melioidosis. Curr Opin Infect Dis 19, 421–8 (2006).
6. Whitmore, A. An account of a glanders-like disease occurring in Rangoon. J
Hyg 13, 1–34 (1913).
7. Whitmore, A. & Krishnaswamy, C.S. An account of the discovery of a
hitherto undescribed infective disease occurring among the population of
Rangoon. Ind. Med. Gaz. 92, 262-7 (1912).
8. Whitmore, A. The bacteriology of an infective disease in Rangoon. Br Med
J 2, 1306–8 (1912).
9. Stanton, A.T. & Fletcher, W. Melioidosis, a new disease of the tropics. Far
Eastern Association of Tropical Medicine: Transactions of the Fourth
Congress (1921).
10. Stanton, A.T. & Fletcher, W. Melioidosis: A disease of rodents
communicable to man. Lancet 205, 10–13 (1925).
11. Stanton, A.T. & Fletcher, W. Melioidosis. (John Bale, Sons & Danielsson:
London, England, 1932).
151

12. Rubin, H.L., Alexander, A.D. & Yager, R.H. Melioidosis--a military medical
problem? Mil Med 128, 538–542 (1963).
13. Howe, C., Sampath, A. & Spotnitz, M. The pseudomallei group: a review. J.
Infect. Dis 124, 598-606 (1971).
14. Cooper, E.B. Melioidosis. JAMA 200, 452-453 (1967).
15. Ognibene, A.J. Medical and infectious diseases in the theater of operations.
Mil Med 152, 14-18 (1987).
16. Weber, D.R., Douglass, L.E., Brundage, W.G. & Stallkamp, T.C. Acute
varieties of melioidosis occurring in U. S. soldiers in Vietnam. Am. J. Med
46, 234-244 (1969).
17. Brundage, W.G., Thuss, C.J. & Walden, D.C. Four fatal cases of melioidosis
in U. S. soldiers in Vietnam. Bacteriologic and pathologic characteristics.
Am. J. Trop. Med. Hyg 17, 183-191 (1968).
18. Patterson, M.C., Darling, C.L. & Blumenthal, J.B. Acute melioidosis in a
soldier home from South Vietnam. JAMA 200, 447-451 (1967).
19. Haynes, W.C. Bergey’s Manual of Determinative Bacteriology. 100 (1957).
20. Miller, W.R., Pannell, L., Cravitz, L., Tanner, W.A. & Ingalls, M.S. Studies
on certain biological characteristics of Malleomyces mallei and
Malleomyces pseudomallei: I. Morphology, cultivation, viability, and
isolation from contaminated specimens. J. Bacteriol 55, 115-126 (1948).
21. Yabuuchi, E. et al. Proposal of Burkholderia gen. nov. and transfer of seven
species of the genus Pseudomonas homology group II to the new genus,
with the type species Burkholderia cepacia (Palleroni and Holmes 1981)
comb. nov. Microbiol. Immunol 36, 1251-1275 (1992).
22. Aubry, P. Mélioïdose. Medecine Tropicale: Diplôme de Médecine Tropicale
des Pays de l’Océan Indien (2010).at
23. Bergey’s Manual of Determinative Bacteriology. (Williams & Wilkins:
Baltimore, Maryland, 1939).
24. Pribram, E. Klassifikation der Schizomyceten (Bakterien): Versuch einer
wissenschaftlichen Klassifikation der Bakterien auf botanischer
Grundlage. (Franz Deuticke: Leipzig und Wien, 1933).
25. Fournier, J. & Chambon, L. La melioidose. Maladie d’actualite et le bacille
de Whitmore. (Editions Medicales Flammarion: Paris, France, 1958).
152

26. Grant, A. & Barwell, C. Chronic melioidosis: a case diagnosed in England.
The Lancet 241, 199–201 (1943).
27. Brindle, C.S. & Cowan, S.T. Flagellation and toxonomy of Whitmore’s
bacillus. J Pathol Bacteriol 63, 571-575 (1951).
28. Redfearn, M.S., Palleroni, N.J. & Stanier, R.Y. A comparative study of
Pseudomonas pseudomallei and Bacillus mallei. J. Gen. Microbiol 43, 293-
313 (1966).
29. Song, H. et al. The early stage of bacterial genome-reductive evolution in
the host. PLoS Pathog 6, e1000922 (2010).
30. Godoy, D. et al. Multilocus sequence typing and evolutionary relationships
among the causative agents of melioidosis and glanders, Burkholderia
pseudomallei and Burkholderia mallei. J. Clin. Microbiol 41, 2068-2079
(2003).
31. Smith, M.D., Wuthiekanun, V., Walsh, A.L. & White, N.J. Quantitative
recovery of Burkholderia pseudomallei from soil in Thailand. Trans. R.
Soc. Trop. Med. Hyg 89, 488-490 (1995).
32. Wuthiekanun, V. et al. Biochemical characteristics of clinical and
environmental isolates of Burkholderia pseudomallei. J. Med. Microbiol
45, 408-412 (1996).
33. Trakulsomboon, S. et al. Epidemiology of arabinose assimilation in
Burkholderia pseudomallei isolated from patients and soil in Thailand.
Southeast Asian J. Trop. Med. Public Health 30, 756-759 (1999).
34. Smith, M.D., Angus, B.J., Wuthiekanun, V. & White, N.J. Arabinose
assimilation defines a nonvirulent biotype of Burkholderia pseudomallei.
Infect. Immun 65, 4319-4321 (1997).
35. Brett, P.J., Deshazer, D. & Woods, D.E. Characterization of Burkholderia
pseudomallei and Burkholderia pseudomallei-like strains. Epidemiol.
Infect 118, 137-148 (1997).
36. Ashdown, L.R. Identification of Pseudomonas pseudomallei in the clinical
laboratory. J. Clin. Pathol 32, 500-504 (1979).
37. DeShazer, D. Virulence of clinical and environmental isolates of
Burkholderia oklahomensis and Burkholderia thailandensis in hamsters
and mice. FEMS Microbiol. Lett 277, 64-69 (2007).
153

38. Brett, P.J., DeShazer, D. & Woods, D.E. Burkholderia thailandensis sp.
nov., a Burkholderia pseudomallei-like species. Int J Syst Bacteriol 48 Pt
1, 317–20 (1998).
39. Pearson, T. et al. Phylogeographic reconstruction of a bacterial species with
high levels of lateral gene transfer. BMC Biol 7, 78 (2009).
40. White, N.J. Melioidosis. Lancet 361, 1715-22 (2003).
41. Wuthiekanun, V. & Peacock, S.J. Management of melioidosis. Expert Rev.
Anti Infect. Ther. 4, 445-455 (2006).
42. Centers for Disease Control and Prevention Biosafety in Microbiological
and Biomedical Laboratories. (National Institutes of Health: Atlanta,
Georgia, 2009).
43. Ashdown, L.R. Melioidosis and safety in the clinical laboratory. J Hosp
Infect 21, 301–6 (1992).
44. Dance, D.A.B., Smith, M.D., Wuthiekanun, V., Walsh, M. & White, N.J.
Melioidosis and laboratory safety. J Hosp Infect 22, 333–334 (1992).
45. Yabuuchi, E. & Arakawa, M. Burkholderia pseudomallei and melioidosis:
be aware in temperate area. Microbiol. Immunol 37, 823-836 (1993).
46. Hamad, M.A. et al. Adaptation and antibiotic tolerance of anaerobic
Burkholderia pseudomallei. Antimicrob Agents Chemother
(2011).doi:10.1128/AAC.00953-10
47. Tong, S., Yang, S., Lu, Z. & He, W. Laboratory investigation of ecological
factors influencing the environmental presence of Burkholderia
pseudomallei. Microbiol. Immunol 40, 451-453 (1996).
48. Wuthiekanun, V., Smith, M.D. & White, N.J. Survival of Burkholderia
pseudomallei in the absence of nutrients. Trans. R. Soc. Trop. Med. Hyg
89, 491 (1995).
49. Moore, R.A., Tuanyok, A. & Woods, D.E. Survival of Burkholderia
pseudomallei in water. BMC Res Notes 1, 11 (2008).
50. Pumpuang, A. et al. Survival of Burkholderia pseudomallei in distilled
water for 16 years. Trans R Soc Trop Med Hyg
(2011).doi:10.1016/j.trstmh.2011.06.004
51. Sheridan, E.A. et al. Evaluation of the Wayson stain for the rapid diagnosis
of melioidosis. J. Clin. Microbiol 45, 1669-1670 (2007).
154

52. Wong, K.T., Puthucheary, S.D. & Vadivelu, J. The histopathology of human
melioidosis. Histopathology 26, 51-55 (1995).
53. Ashdown, L. An improved screening technique for isolation of
Pseudomonas pseudomallei from clinical specimens. Pathology 11, 293-97
(1979).
54. Chantratita, N. et al. Biological relevance of colony morphology and
phenotypic switching by Burkholderia pseudomallei. J. Bacteriol 189, 807-
817 (2007).
55. Peacock, S.J. et al. Comparison of Ashdown’s medium, Burkholderia
cepacia medium, and Burkholderia pseudomallei selective agar for clinical
isolation of Burkholderia pseudomallei. J. Clin. Microbiol 43, 5359-5361
(2005).
56. Glass, M.B., Beesley, C.A., Wilkins, P.P. & Hoffmaster, A.R. Comparison of
four selective media for the isolation of Burkholderia mallei and
Burkholderia pseudomallei. Am. J. Trop. Med. Hyg 80, 1023-1028 (2009).
57. Amornchai, P. et al. Accuracy of Burkholderia pseudomallei identification
using the API 20NE system and a latex agglutination test. J. Clin. Microbiol
45, 3774-3776 (2007).
58. Lowe, P., Engler, C. & Norton, R. Comparison of automated and
nonautomated systems for identification of Burkholderia pseudomallei. J.
Clin. Microbiol 40, 4625-4627 (2002).
59. Lowe, P., Haswell, H. & Lewis, K. Use of various common isolation media
to evaluate the new VITEK 2 colorimetric GN Card for identification of
Burkholderia pseudomallei. J. Clin. Microbiol 44, 854-856 (2006).
60. Walsh, A.L. & Wuthiekanun, V. The laboratory diagnosis of melioidosis. Br.
J. Biomed. Sci 53, 249-253 (1996).
61. Wuthiekanun, V., Anuntagool, N., White, N.J. & Sirisinha, S. Short report:
a rapid method for the differentiation of Burkholderia pseudomallei and
Burkholderia thailandensis. Am. J. Trop. Med. Hyg 66, 759-761 (2002).
62. Wuthiekanun, V. et al. Quantitation of B. pseudomallei in clinical samples.
Am. J. Trop. Med. Hyg 77, 812-813 (2007).
63. Walsh, A.L. et al. Immunofluorescence microscopy for the rapid diagnosis
of melioidosis. J. Clin. Pathol 47, 377-379 (1994).
155

64. Wuthiekanun, V. et al. Rapid immunofluorescence microscopy for
diagnosis of melioidosis. Clin. Diagn. Lab. Immunol 12, 555-556 (2005).
65. Walsh, A.L. et al. Prognostic significance of quantitative bacteremia in
septicemic melioidosis. Clin. Infect. Dis 21, 1498-1500 (1995).
66. Ruppitsch, W. et al. Suitability of partial 16S ribosomal RNA gene sequence
analysis for the identification of dangerous bacterial pathogens. J. Appl.
Microbiol 102, 852-859 (2007).
67. Tomaso, H. et al. Development of 5’ nuclease real-time PCR assays for the
rapid identification of the Burkholderia mallei/Burkholderia pseudomallei
complex. Diagn. Mol. Pathol 13, 247-253 (2004).
68. Wattiau, P. et al. Identification of Burkholderia pseudomallei and related
bacteria by multiple-locus sequence typing-derived PCR and real-time
PCR. J. Clin. Microbiol 45, 1045-1048 (2007).
69. Thibault, F.M., Valade, E. & Vidal, D.R. Identification and discrimination
of Burkholderia pseudomallei, B. mallei, and B. thailandensis by real-time
PCR targeting type III secretion system genes. J. Clin. Microbiol 42, 5871-
5874 (2004).
70. Lee, M.-A., Wang, D. & Yap, E.H. Detection and differentiation of
Burkholderia pseudomallei, Burkholderia mallei and Burkholderia
thailandensis by multiplex PCR. FEMS Immunol. Med. Microbiol 43, 413-
417 (2005).
71. Antonov, V.A. et al. Molecular identification and typing of Burkholderia
pseudomallei and Burkholderia mallei: when is enough enough? Trans. R.
Soc. Trop. Med. Hyg 102 Suppl 1, S134-139 (2008).
72. Chantratita, N. et al. Loop-mediated isothermal amplification method
targeting the TTS1 gene cluster for detection of Burkholderia pseudomallei
and diagnosis of melioidosis. J Clin Microbiol 46, 568-73 (2008).
73. Lertpatanasuwan, N. et al. Arabinose-positive Burkholderia pseudomallei
infection in humans: case report. Clin. Infect. Dis 28, 927-928 (1999).
74. Glass, M.B. et al. Pneumonia and septicemia caused by Burkholderia
thailandensis in the United States. J. Clin. Microbiol 44, 4601-4604
(2006).
75. Srinivasan, A. et al. Glanders in a military research microbiologist. N. Engl.
J. Med 345, 256-258 (2001).
156

76. Kenny, D.J., Russell, P., Rogers, D., Eley, S.M. & Titball, R.W. In vitro
susceptibilities of Burkholderia mallei in comparison to those of other
pathogenic Burkholderia spp. Antimicrob. Agents Chemother 43, 2773-
2775 (1999).
77. Thibault, F.M., Hernandez, E., Vidal, D.R., Girardet, M. & Cavallo, J.-D.
Antibiotic susceptibility of 65 isolates of Burkholderia pseudomallei and
Burkholderia mallei to 35 antimicrobial agents. J. Antimicrob. Chemother.
54, 1134-1138 (2004).
78. Wernery, U. et al. Natural Burkholderia mallei infection in Dromedary,
Bahrain. Emerging Infect. Dis. 17, 1277-1279 (2011).
79. Pruksachartvuthi, S., Aswapokee, N. & Thankerngpol, K. Survival of
Pseudomonas pseudomallei in human phagocytes. J. Med. Microbiol 31,
109-114 (1990).
80. Jones, A.L., Beveridge, T.J. & Woods, D.E. Intracellular survival of
Burkholderia pseudomallei. Infect. Immun 64, 782-790 (1996).
81. French, C.T. et al. Dissection of the Burkholderia intracellular life cycle
using a photothermal nanoblade. Proc. Natl. Acad. Sci. U.S.A 108, 12095-
12100 (2011).
82. Kespichayawattana, W., Rattanachetkul, S., Wanun, T., Utaisincharoen, P.
& Sirisinha, S. Burkholderia pseudomallei induces cell fusion and actin-
associated membrane protrusion: a possible mechanism for cell-to-cell
spreading. Infect. Immun 68, 5377-5384 (2000).
83. Breitbach, K. et al. Actin-based motility of Burkholderia pseudomallei
involves the Arp 2/3 complex, but not N-WASP and Ena/VASP proteins.
Cell. Microbiol 5, 385-393 (2003).
84. Stevens, M.P. et al. Identification of a bacterial factor required for actin-
based motility of Burkholderia pseudomallei. Mol. Microbiol 56, 40-53
(2005).
85. Norville, I.H. et al. A Burkholderia pseudomallei Mip-like protein has
rapamycin inhibitable peptidyl-prolyl isomerase activity and has
pleiotropic effects on virulence. Infect Immun
(2011).doi:10.1128/IAI.00134-11
86. Reckseidler, S.L., DeShazer, D., Sokol, P.A. & Woods, D.E. Detection of
bacterial virulence genes by subtractive hybridization: identification of
157

capsular polysaccharide of Burkholderia pseudomallei as a major virulence
determinant. Infect. Immun 69, 34-44 (2001).
87. Atkins, T. et al. Characterisation of an acapsular mutant of Burkholderia
pseudomallei identified by signature tagged mutagenesis. J. Med.
Microbiol 51, 539-547 (2002).
88. Reckseidler-Zenteno, S.L., DeVinney, R. & Woods, D.E. The capsular
polysaccharide of Burkholderia pseudomallei contributes to survival in
serum by reducing complement factor C3b deposition. Infect. Immun 73,
1106-1115 (2005).
89. Warawa, J.M., Long, D., Rosenke, R., Gardner, D. & Gherardini, F.C. Role
for the Burkholderia pseudomallei capsular polysaccharide encoded by the
wcb operon in acute disseminated melioidosis. Infect. Immun 77, 5252-
5261 (2009).
90. Wikraiphat, C. et al. Comparative in vivo and in vitro analyses of putative
virulence factors of Burkholderia pseudomallei using lipopolysaccharide,
capsule and flagellin mutants. FEMS Immunol. Med. Microbiol 56, 253-
259 (2009).
91. Reckseidler-Zenteno, S.L. et al. Characterization of the type III capsular
polysaccharide produced by Burkholderia pseudomallei. J. Med. Microbiol
59, 1403-1414 (2010).
92. Brett, P.J., Mah, D.C. & Woods, D.E. Isolation and characterization of
Pseudomonas pseudomallei flagellin proteins. Infect. Immun 62, 1914-1919
(1994).
93. DeShazer, D., Brett, P.J., Carlyon, R. & Woods, D.E. Mutagenesis of
Burkholderia pseudomallei with Tn5-OT182: isolation of motility mutants
and molecular characterization of the flagellin structural gene. J. Bacteriol
179, 2116-2125 (1997).
94. Chua, K.L., Chan, Y.Y. & Gan, Y.H. Flagella are virulence determinants of
Burkholderia pseudomallei. Infect. Immun 71, 1622-1629 (2003).
95. Inglis, T.J.J., Robertson, T., Woods, D.E., Dutton, N. & Chang, B.J.
Flagellum-mediated adhesion by Burkholderia pseudomallei precedes
invasion of Acanthamoeba astronyxis. Infect. Immun 71, 2280-2282
(2003).
158

96. Chuaygud, T., Tungpradabkul, S., Sirisinha, S., Chua, K.L. &
Utaisincharoen, P. A role of Burkholderia pseudomallei flagella as a
virulent factor. Trans. R. Soc. Trop. Med. Hyg 102 Suppl 1, S140-144
(2008).
97. Matsuura, M., Kawahara, K., Ezaki, T. & Nakano, M. Biological activities of
lipopolysaccharide of Burkholderia (Pseudomonas) pseudomallei. FEMS
Microbiol. Lett 137, 79-83 (1996).
98. DeShazer, D., Brett, P.J. & Woods, D.E. The type II O-antigenic
polysaccharide moiety of Burkholderia pseudomallei lipopolysaccharide is
required for serum resistance and virulence. Mol. Microbiol 30, 1081-1100
(1998).
99. Novem, V. et al. Structural and biological diversity of lipopolysaccharides
from Burkholderia pseudomallei and Burkholderia thailandensis. Clin.
Vaccine Immunol (2009).doi:10.1128/CVI.00472-08
100. Ulrich, R.L. et al. Role of quorum sensing in the pathogenicity of
Burkholderia pseudomallei. J. Med. Microbiol 53, 1053-1064 (2004).
101. Juhas, M. et al. Global regulation of quorum sensing and virulence by VqsR
in Pseudomonas aeruginosa. Microbiology (Reading, Engl.) 150, 831-841
(2004).
102. Valade, E. et al. The PmlI-PmlR quorum-sensing system in Burkholderia
pseudomallei plays a key role in virulence and modulates production of the
MprA protease. J. Bacteriol 186, 2288-2294 (2004).
103. Song, Y., Xie, C., Ong, Y.-M., Gan, Y.-H. & Chua, K.-L. The BpsIR quorum-
sensing system of Burkholderia pseudomallei. J. Bacteriol 187, 785-790
(2005).
104. Diggle, S.P. et al. Functional genetic analysis reveals a 2-Alkyl-4-quinolone
signaling system in the human pathogen Burkholderia pseudomallei and
related bacteria. Chem. Biol 13, 701-710 (2006).
105. Lumjiaktase, P. et al. Quorum sensing regulates dpsA and the oxidative
stress response in Burkholderia pseudomallei. Microbiology (Reading,
Engl.) 152, 3651-3659 (2006).
106. Vanaporn, M. et al. Superoxide dismutase C is required for intracellular
survival and virulence of Burkholderia pseudomallei. Microbiology
(Reading, Engl.) 157, 2392-2400 (2011).
159

107. Essex-Lopresti, A.E. et al. A type IV pilin, PilA, contributes to adherence of
Burkholderia pseudomallei and virulence in vivo. Infect. Immun 73, 1260-
1264 (2005).
108. Stevens, M.P. et al. An Inv/Mxi-Spa-like type III protein secretion system
in Burkholderia pseudomallei modulates intracellular behaviour of the
pathogen. Mol. Microbiol 46, 649-659 (2002).
109. Stevens, M.P. et al. Attenuated virulence and protective efficacy of a
Burkholderia pseudomallei bsa type III secretion mutant in murine models
of melioidosis. Microbiology (Reading, Engl.) 150, 2669-2676 (2004).
110. Warawa, J. & Woods, D.E. Type III secretion system cluster 3 is required
for maximal virulence of Burkholderia pseudomallei in a hamster infection
model. FEMS Microbiol. Lett 242, 101-108 (2005).
111. Pilatz, S. et al. Identification of Burkholderia pseudomallei genes required
for the intracellular life cycle and in vivo virulence. Infect. Immun 74,
3576-3586 (2006).
112. Burtnick, M.N. et al. Burkholderia pseudomallei type III secretion system
mutants exhibit delayed vacuolar escape phenotypes in RAW 264.7 murine
macrophages. Infect. Immun 76, 2991-3000 (2008).
113. Sun, G.W. et al. Identification of a regulatory cascade controlling Type III
Secretion System 3 gene expression in Burkholderia pseudomallei. Mol.
Microbiol 76, 677-689 (2010).
114. Schell, M.A. et al. Type VI secretion is a major virulence determinant in
Burkholderia mallei. Mol. Microbiol 64, 1466-1485 (2007).
115. Shalom, G., Shaw, J.G. & Thomas, M.S. In vivo expression technology
identifies a type VI secretion system locus in Burkholderia pseudomallei
that is induced upon invasion of macrophages. Microbiology (Reading,
Engl.) 153, 2689-2699 (2007).
116. Burtnick, M.N. et al. The cluster 1 type VI secretion system is a major
virulence determinant in Burkholderia pseudomallei. Infect. Immun 79,
1512-1525 (2011).
117. Ralph, P. & Nakoinz, I. Antibody-dependent killing of erythrocyte and
tumor targets by macrophage-related cell lines: enhancement by PPD and
LPS. J. Immunol 119, 950-954 (1977).
160

118. Inglis, T.J. et al. Interaction between Burkholderia pseudomallei and
Acanthamoeba species results in coiling phagocytosis, endamebic bacterial
survival, and escape. Infect. Immun 68, 1681-1686 (2000).
119. Utaisincharoen, P., Tangthawornchaikul, N., Kespichayawattana, W.,
Chaisuriya, P. & Sirisinha, S. Burkholderia pseudomallei interferes with
inducible nitric oxide synthase (iNOS) production: a possible mechanism of
evading macrophage killing. Microbiol. Immunol 45, 307-313 (2001).
120. Dance, D.A. Melioidosis: the tip of the iceberg? Clin. Microbiol. Rev 4, 52-
60 (1991).
121. Dance, D.A. Melioidosis as an emerging global problem. Acta Trop 74, 115-
119 (2000).
122. Gambier, A. Un cas de mélioïdose observé à Phôm-Penh. Bulletins de la
Société de pathologie exotique et de ses filiales de l’Ouest africain et de
Madagascar 23, 436–441 (1930).
123. Paton, J.P.J., Peck, C.R. & van de Schaff, A. Report on a case of melioidosis
from Siam. Br Med J 1, 336–337 (1947).
124. Chittivej, C., Buspavanij, S. & Chaovanasai, A. Melioidosis with case report
in a Thai. Royal Thai Army Medical Journal 68, 11–17 (1955).
125. Chaowagul, W. et al. Melioidosis: a major cause of community-acquired
septicemia in northeastern Thailand. J. Infect. Dis 159, 890-899 (1989).
126. Punyagupta, S. Melioidosis. Review of 686 cases and presentation of a new
clinical classification. Melioidosis 217–229 (1989).
127. Krishnaswami, C.S. Morphia injectors’ septicaemia. Ind. Med. Gaz. 52,
296–299 (1917).
128. Aung, D.M.K. Melioidosis—a hidden disease in Myanmar. Myanmar
journal of current medical practice 5, 57–59 (2000).
129. Harries, E.J., Lewis, A.A.G., Waring, J.W.B. & Dowling, E.J. Melioidosis
treated with sulphonamides and penicillin. Lancet 251, 363–366 (1948).
130. Lee, S.C., Ling, T.S., Chen, J.C., Huang, B.Y. & Sheih, W.B. Melioidosis with
adrenal gland abscess. Am. J. Trop. Med. Hyg 61, 34-36 (1999).
131. Wuthiekanun, V. et al. Short report: Melioidosis in Myanmar: forgotten but
not gone? Am. J. Trop. Med. Hyg 75, 945-946 (2006).
132. Cheng, A.C. & Currie, B.J. Melioidosis: epidemiology, pathophysiology, and
management. Clin Microbiol Rev 18, 383-416 (2005).
161

133. Currie, B.J., Dance, D.A.B. & Cheng, A.C. The global distribution of
Burkholderia pseudomallei and melioidosis: an update. Trans. R. Soc.
Trop. Med. Hyg 102 Suppl 1, S1-4 (2008).
134. Antony, B. et al. Spectrum of melioidosis in the suburbs of Mangalore, S
West Coast of India. Southeast Asian J. Trop. Med. Public Health 41, 169-
174 (2010).
135. Saravu, K. et al. Melioidosis in southern India: epidemiological and clinical
profile. Southeast Asian J. Trop. Med. Public Health 41, 401-409 (2010).
136. Mukhopadhyay, C. et al. Molecular Characterization of Clinical
Burkholderia pseudomallei Isolates from India. Am. J. Trop. Med. Hyg 85,
121-123 (2011).
137. O’Sullivan, B.P. et al. Burkholderia pseudomallei infection in a child with
cystic fibrosis: acquisition in the Western hemisphere. Chest 140, 239-242
(2011).
138. Cuadros, J. et al. Case report: melioidosis imported from West Africa to
Europe. Am. J. Trop. Med. Hyg. 85, 282-284 (2011).
139. Strauss, J.M. et al. Melioidosis in Malaysia. IV. Intensive ecological study
of Carey Island, Selangor, for Pseudomonas pseudomallei. Med J Malaya
24, 94-100 (1969).
140. Chambon, L. Isolement du Bacille de Whitmore a partir du milieu
exterieur. Ann Inst Pasteur (Paris) 89, 229-235 (1955).
141. Krauss, H. Zoonoses: infectious diseases transmissible from animals to
humans. (ASM Press: 2003).
142. Shakespeare, M. Zoonoses. (Pharmaceutical Press: 2002).
143. Choy, J.L., Mayo, M., Janmaat, A. & Currie, B.J. Animal melioidosis in
Australia. Acta Tropica 74, 153–158 (2000).
144. Yap, E.H. et al. Comparison of Pseudomonas pseudomallei from humans,
animals, soil and water by restriction endonuclease analysis. Singapore
Med J 36, 60-62 (1995).
145. Mollaret, H.H. «L’affaire du Jardin des plantes» ou comment le mélioïdose
fit son apparition en France. Med Mal Infect 18, 643–54 (1988).
146. Baharsefat, M. & Amjadi, A.R. Equine melioidosis in Iran. Arch Inst Razi
22, 209–213 (1970).
162

147. Galimand, M. & Dodin, A. Le point sur la melioidose dans le monde. Bull
Soc Pathol Exot Filiales 75, 375-383 (1982).
148. Dodin, A. & Galimand, M. Naissance, vie et recul d’une maladie infectieuse,
la melioidose en pays tempere. Arch Inst Pasteur Tunis 63, 69-73 (1986).
149. Sprague, L.D. & Neubauer, H. Melioidosis in animals: a review on
epizootiology, diagnosis and clinical presentation. J. Vet. Med. B Infect.
Dis. Vet. Public Health 51, 305-320 (2004).
150. Woods, D.E., Jones, A.L. & Hill, P.J. Interaction of insulin with
Pseudomonas pseudomallei. Infect Immun 61, 4045-50 (1993).
151. Wiersinga, W.J. et al. Toll-like receptor 2 impairs host defense in Gram-
negative sepsis caused by Burkholderia pseudomallei (melioidosis). PLoS
Med 4, e248 (2007).
152. Tuanyok, A., Tom, M., Dunbar, J. & Woods, D.E. Genome-wide expression
analysis of Burkholderia pseudomallei infection in a hamster model of
acute melioidosis. Infect. Immun 74, 5465-5476 (2006).
153. Dannenberg, A.M. & Scott, E.M. Melioidosis: pathogenesis and immunity
in mice and hamsters. III. Effect of vaccination with avirulent strains of
Pseudomonas pseudomallei on the resistance to the establishment and the
resistance to the progress of respiratory melioidosis caused by virulent
strains; all-or-none aspects of this disease. J. Immunol 84, 233-246
(1960).
154. Miller, W.R., Pannell, L., Cravitz, L., Tanner, W.A. & Rosebury, T. Studies
on certain biological characteristics of Malleomyces mallei and
Malleomyces pseudomallei: II. Virulence and infectivity for animals. J.
Bacteriol 55, 127-135 (1948).
155. Strauss, J.M., Groves, M.G., Mariappan, M. & Ellison, D.W. Melioidosis in
Malaysia. II. Distribution of Pseudomonas pseudomallei in soil and surface
water. Am. J. Trop. Med. Hyg 18, 698-702 (1969).
156. Strauss, J.M., Alexander, A.D., Rapmund, G., Gan, E. & Dorsey, A.E.
Melioidosis in Malaysia. III. Antibodies to Pseudomonas pseudomallei in
the human population. Am. J. Trop. Med. Hyg 18, 703-707 (1969).
157. Vuddhakul, V. et al. Epidemiology of Burkholderia pseudomallei in
Thailand. Am. J. Trop. Med. Hyg 60, 458-461 (1999).
163

158. Mayo, M. et al. Burkholderia pseudomallei in unchlorinated domestic bore
water, tropical northern Australia. Emerging Infect. Dis 17, 1283-1285
(2011).
159. Chen, Y.-S. et al. Distribution of melioidosis cases and viable Burkholderia
pseudomallei in soil: evidence for emerging melioidosis in Taiwan. J. Clin.
Microbiol 48, 1432-1434 (2010).
160. Limmathurotsakul, D. et al. Burkholderia pseudomallei is spatially
distributed in soil in northeast Thailand. PLoS Negl Trop Dis 4, e694
(2010).
161. Lin, H.-H. et al. Burkholderia multivorans acts as an antagonist against the
growth of Burkholderia pseudomallei in soil. Microbiol Immunol
(2011).doi:10.1111/j.1348-0421.2011.00365.x
162. Kaestli, M. et al. Landscape changes influence the occurrence of the
melioidosis bacterium Burkholderia pseudomallei in soil in northern
Australia. PLoS Negl Trop Dis 3, e364 (2009).
163. Inglis, T.J.J. & Sagripanti, J.-L. Environmental factors that affect the
survival and persistence of Burkholderia pseudomallei. Appl. Environ.
Microbiol 72, 6865-6875 (2006).
164. Marolda, C.L., Hauröder, B., John, M.A., Michel, R. & Valvano, M.A.
Intracellular survival and saprophytic growth of isolates from the
Burkholderia cepacia complex in free-living amoebae. Microbiology
(Reading, Engl.) 145 ( Pt 7), 1509-1517 (1999).
165. Howard, K. & Inglis, T.J.J. Disinfection of Burkholderia pseudomallei in
potable water. Water Res 39, 1085-1092 (2005).
166. Currie, B.J. & Jacups, S.P. Intensity of rainfall and severity of melioidosis,
Australia. Emerging Infect. Dis 9, 1538-1542 (2003).
167. Lo, T.J., Ang, L.W., James, L. & Goh, K.T. Melioidosis in a tropical city
state, Singapore. Emerging Infect. Dis 15, 1645-1647 (2009).
168. Suputtamongkol, Y. et al. The epidemiology of melioidosis in Ubon
Ratchatani, northeast Thailand. Int J Epidemiol 23, 1082-1090 (1994).
169. Limmathurotsakul, D. et al. Increasing incidence of human melioidosis in
Northeast Thailand. Am J Trop Med Hyg 82, 1113-1117 (2010).
170. Sam, I.-C. & Puthucheary, S.D. Melioidosis and rainfall in Kuala Lumpur,
Malaysia. J. Infect 54, 519-520 (2007).
164

171. Cheng, A.C., Jacups, S.P., Gal, D., Mayo, M. & Currie, B.J. Extreme weather
events and environmental contamination are associated with case-clusters
of melioidosis in the Northern Territory of Australia. Int J Epidemiol 35,
323-329 (2006).
172. Su, H.-P., Chan, T.-C. & Chang, C.-C. Typhoon-related leptospirosis and
melioidosis, Taiwan, 2009. Emerging Infect. Dis 17, 1322-1324 (2011).
173. Lim, M.K., Tan, E.H., Soh, C.S. & Chang, T.L. Burkholderia pseudomallei
infection in the Singapore Armed Forces from 1987 to 1994—an
epidemiological review. Ann. Acad. Med. Singap 26, 13-17 (1997).
174. Suputtamongkol, Y. et al. Risk factors for melioidosis and bacteremic
melioidosis. Clin. Infect. Dis 29, 408-413 (1999).
175. Currie, B.J., Ward, L. & Cheng, A.C. The epidemiology and clinical
spectrum of melioidosis: 540 cases from the 20 year Darwin prospective
study. PLoS Negl Trop Dis 4, e900 (2010).
176. Currie, B.J. et al. The epidemiology of melioidosis in Australia and Papua
New Guinea. Acta Trop 74, 121-127 (2000).
177. Tan, A.L., Ang, B.S. & Ong, Y.Y. Melioidosis: epidemiology and antibiogram
of cases in Singapore. Singapore Med J 31, 335-337 (1990).
178. Heng, B.H., Sun, Y., Cheah, J.T.S. & Jong, M. The Singapore National
Healthcare Group Diabetes Registry—descriptive epidemiology of type 2
diabetes mellitus. Ann. Acad. Med. Singap 39, 348-352 (2010).
179. Chierakul, W. et al. Short report: disease severity and outcome of
melioidosis in HIV coinfected individuals. Am. J. Trop. Med. Hyg 73, 1165-
1166 (2005).
180. Mackowiak, P.A. & Smith, J.W. Septicemic melioidosis. Occurrence
following acute influenza A six years after exposure in Vietnam. JAMA
240, 764-766 (1978).
181. Tan, G.-Y.G. et al. Burkholderia pseudomallei aerosol infection results in
differential inflammatory responses in BALB/c and C57Bl/6 mice. J Med
Microbiol 57, 508-515 (2008).
182. Green, R.N. & Tuffnell, P.G. Laboratory acquired melioidosis. Am. J. Med
44, 599-605 (1968).
165

183. Ho, P.L. et al. Activity of five fluoroquinolones against 71 isolates of
Burkholderia pseudomallei. J. Antimicrob. Chemother. 49, 1042 -1044
(2002).
184. Ralph, A., McBride, J. & Currie, B.J. Transmission of Burkholderia
pseudomallei via breast milk in northern Australia. Pediatr. Infect. Dis. J
23, 1169-1171 (2004).
185. West, T.E. et al. Pathogenicity of high-dose enteral inoculation of
Burkholderia pseudomallei to mice. Am. J. Trop. Med. Hyg 83, 1066-1069
(2010).
186. Currie, B.J. et al. A cluster of melioidosis cases from an endemic region is
clonal and is linked to the water supply using molecular typing of
Burkholderia pseudomallei isolates. Am. J. Trop. Med. Hyg 65, 177-179
(2001).
187. Inglis, T.J. et al. Acute melioidosis outbreak in Western Australia.
Epidemiol. Infect 123, 437-443 (1999).
188. Inglis, T.J. et al. Burkholderia pseudomallei traced to water treatment
plant in Australia. Emerging Infect. Dis 6, 56-59 (2000).
189. Chierakul, W. et al. Melioidosis in 6 tsunami survivors in southern
Thailand. Clin. Infect. Dis 41, 982-990 (2005).
190. Wuthiekanun, V. et al. Serological evidence for increased human exposure
to Burkholderia pseudomallei following the tsunami in southern Thailand.
J. Clin. Microbiol 44, 239-240 (2006).
191. Athan, E., Allworth, A.M., Engler, C., Bastian, I. & Cheng, A.C. Melioidosis
in tsunami survivors. Emerging Infect. Dis 11, 1638-1639 (2005).
192. Othman, N. et al. Infections in post-tsunami victims. Pediatr. Infect. Dis. J
26, 960-961 (2007).
193. Ciervo, A., Mattei, R. & Cassone, A. Melioidosis in an Italian tourist injured
by the tsunami in Thailand. J Chemother 18, 443-444 (2006).
194. Svensson, E., Welinder-Olsson, C., Claesson, B.A. & Studahl, M. Cutaneous
melioidosis in a Swedish tourist after the tsunami in 2004. Scand. J. Infect.
Dis 38, 71-74 (2006).
195. Allworth, A.M. Tsunami lung: a necrotising pneumonia in survivors of the
Asian tsunami. Med. J. Aust 182, 364 (2005).
166

196. McCormick, J.B. et al. Human-to-human transmission of Pseudomonas
pseudomallei. Ann. Intern. Med 83, 512-513 (1975).
197. Kunakorn, M., Jayanetra, P. & Tanphaichitra, D. Man-to-man transmission
of melioidosis. Lancet 337, 1290-1291 (1991).
198. Abbink, F.C., Orendi, J.M. & de Beaufort, A.J. Mother-to-child
transmission of Burkholderia pseudomallei. N. Engl. J. Med 344, 1171-1172
(2001).
199. Simpson, A.J. et al. Comparison of imipenem and ceftazidime as therapy
for severe melioidosis. Clin. Infect. Dis 29, 381-387 (1999).
200. Chetchotisakd, P., Porramatikul, S., Mootsikapun, P., Anunnatsiri, S. &
Thinkhamrop, B. Randomized, double-blind, controlled study of
cefoperazone-sulbactam plus cotrimoxazole versus ceftazidime plus
cotrimoxazole for the treatment of severe melioidosis. Clin. Infect. Dis 33,
29–34 (2001).
201. Apisarnthanarak, A., Apisarnthanarak, P. & Mundy, L.M. Computed
tomography characteristics of Burkholderia pseudomallei liver abscess.
Clin. Infect. Dis 42, 989-993 (2006).
202. Teparrakkul, P. et al. Rheumatological manifestations in patients with
melioidosis. Southeast Asian J. Trop. Med. Public Health 39, 649-655
(2008).
203. Siripanthong, S. et al. Corneal ulcer caused by Pseudomonas pseudomallei:
report of three cases. Rev. Infect. Dis 13, 335-337 (1991).
204. Chen, K.-J., Sun, M.-H., Hou, C.-H., Sun, C.-C. & Chen, T.-L. Burkholderia
pseudomallei endophthalmitis. J. Clin. Microbiol 45, 4073-4074 (2007).
205. Yang, I.-H., Lee, J.-J., Liu, J.-W. & Teng, M.-C. Melioidosis with
endophthalmitis. Arch. Ophthalmol 124, 1501-1502 (2006).
206. Demar, M., Ferroni, A., Dupont, B., Eliaszewicz, M. & Bourée, P.
Suppurative epididymo-orchitis and chronic prostatitis caused by
Burkholderia pseudomallei: a case report and review. J Travel Med 12,
108-112 (2005).
207. Koh, K.B. Melioidosis presenting as epididymo-orchitis. Singapore Med J
36, 446 (1995).
167

208. Anunnatsiri, S., Chetchotisakd, P. & Kularbkaew, C. Mycotic aneurysm in
Northeast Thailand: the importance of Burkholderia pseudomallei as a
causative pathogen. Clin. Infect. Dis 47, 1436-1439 (2008).
209. Currie, B.J., Fisher, D.A., Howard, D.M. & Burrow, J.N. Neurological
melioidosis. Acta Trop 74, 145-151 (2000).
210. Limmathurotsakul, D. et al. Variable presentation of neurological
melioidosis in Northeast Thailand. Am. J. Trop. Med. Hyg 77, 118-120
(2007).
211. Kasantikul, V., Lerdlum, S. & Suwanwela, N. Cerebral abscesses due to
Pseudomonas pseudomallei. J Med Assoc Thai 75, 536-541 (1992).
212. Chadwick, D.R., Ang, B., Sitoh, Y.Y. & Lee, C.C. Cerebral melioidosis in
Singapore: a review of five cases. Trans. R. Soc. Trop. Med. Hyg 96, 72-76
(2002).
213. Chlebicki, M.P., Kurup, A. & Sin, Y.K. Burkholderia pseudomallei
meningitis following inadequate treatment of melioidotic mycotic
aneurysm. Singapore Med J 49, e219-221 (2008).
214. Currie, B.J. et al. Endemic melioidosis in tropical northern Australia: a 10-
year prospective study and review of the literature. Clin. Infect. Dis 31, 981-
986 (2000).
215. Mayer, J.H. & Finlayson, M.H. Chronic melioidosis. S. Afr. Med. J. XVIII,
109–112 (1944).
216. Falade, O.O., Antonarakis, E.S., Kaul, D.R., Saint, S. & Murphy, P.A.
Clinical problem-solving. Beware of first impressions. N Engl J Med 359,
628-34 (2008).
217. Vidyalakshmi, K. et al. Tuberculosis mimicked by melioidosis. Int. J.
Tuberc. Lung Dis 12, 1209-1215 (2008).
218. Chetchotisakd, P., Anunnatsiri, S., Kiatchoosakun, S. & Kularbkaew, C.
Melioidosis pericarditis mimicking tuberculous pericarditis. Clin. Infect.
Dis 51, e46-49 (2010).
219. Saravu, K. et al. Melioidosis—a case series from south India. Trans R Soc
Trop Med Hyg 102 Suppl 1, S18-20 (2008).
220. Currie, B.J., Fisher, D.A., Anstey, N.M. & Jacups, S.P. Melioidosis: acute
and chronic disease, relapse and re-activation. Trans. R. Soc. Trop. Med.
Hyg 94, 301-304 (2000).
168

221. Pongrithsukda, V., Simakachorn, N. & Pimda, J. Childhood melioidosis in
northeastern Thailand. Southeast Asian J. Trop. Med. Public Health 19,
309-316 (1988).
222. Dance, D.A. et al. Acute suppurative parotitis caused by Pseudomonas
pseudomallei in children. J. Infect. Dis 159, 654-660 (1989).
223. Lumbiganon, P. & Viengnondha, S. Clinical manifestations of melioidosis
in children. Pediatr. Infect. Dis. J 14, 136-140 (1995).
224. Pagnarith, Y. et al. Emergence of pediatric melioidosis in Siem Reap,
Cambodia. Am. J. Trop. Med. Hyg 82, 1106-1112 (2010).
225. Sam, I.-C. & Puthucheary, S.D. Melioidosis in children from Kuala Lumpur,
Malaysia. Ann Trop Paediatr 26, 219-224 (2006).
226. Kandasamy, Y. & Norton, R. Paediatric melioidosis in North Queensland,
Australia. J Paediatr Child Health 44, 706-708 (2008).
227. Kandasamy, Y. & Somasundaram, P. Paediatric melioidosis with septic
shock in a previously-well child. Singapore Med J 48, e109-110 (2007).
228. Holland, D.J., Wesley, A., Drinkovic, D. & Currie, B.J. Cystic fibrosis and
Burkholderia pseudomallei infection: an emerging problem? Clin. Infect.
Dis. 35, e138 -e140 (2002).
229. Gilbert, D.N., Moore, W.L., Jr, Hedberg, C.L. & Sanford, J.P. Potential
medical problems in personnel returning from Vietnam. Ann. Intern. Med
68, 662-678 (1968).
230. Sanford, J.P. & Moore, W.L., Jr Recrudescent melioidosis: a southeast
asian legacy. Am. Rev. Respir. Dis 104, 452-453 (1971).
231. Sanford, J.P. Melioidosis: forgotten but not gone. Trans. Am. Clin.
Climatol. Assoc 89, 201-205 (1978).
232. Spotnitz, M. Disease may be Vietnamese time bomb. Med World News 7,
55 (1966).
233. Koponen, M.A., Zlock, D., Palmer, D.L. & Merlin, T.L. Melioidosis.
Forgotten, but not gone! Arch. Intern. Med 151, 605-608 (1991).
234. Mays, E.E. & Ricketts, E.A. Melioidosis: recrudescence associated with
bronchogenic carcinoma twenty-six years following initial geographic
exposure. Chest 68, 261-263 (1975).
235. Clayton, A.J., Lisella, R.S. & Martin, D.G. Melioidosis: a serological survey
in military personnel. Mil Med 138, 24-26 (1973).
169

236. Phung, L.V., Quynh, H.T., Yabuuchi, E. & Dance, D.A. Pilot study of
exposure to Pseudomonas pseudomallei in northern Vietnam. Trans. R.
Soc. Trop. Med. Hyg 87, 416 (1993).
237. Phung, L.V., Dung, D.H., Thu, P.M., Hong, N.H. & Dance, D.A.B. Exposure
to Pseudomonas pseudomallei in Northern Vietnam. Melioidosis:
Prevailing Problems and Future Directions 25 (1994).
238. Wuthiekanun, V., Suputtamongkol, Y., Simpson, A.J., Kanaphun, P. &
White, N.J. Value of throat swab in diagnosis of melioidosis. J. Clin.
Microbiol 39, 3801-3802 (2001).
239. Schülin, T. & Steinmetz, I. Chronic melioidosis in a patient with cystic
fibrosis. J. Clin. Microbiol 39, 1676-1677 (2001).
240. Currie, B.J. Melioidosis: an important cause of pneumonia in residents of
and travellers returned from endemic regions. Eur. Respir. J 22, 542-550
(2003).
241. de Moor, C.E., Soekarnen & van der Walle, N. Melioidosis auf Java.
Mededelingen van den Dienst der Volksgezondheid in Nederlands Indie
21, 206–222, 226. (1932).
242. Brygoo, E.R. Contribution à l‘étude des agglutinines naturelles pour le
bacille de Whitmore. Les agglutinines spécifiques dans quelques cas de
mélioïdose. Bull Soc Pathol Exot Filiales 46, 347-353 (1953).
243. Wuthiekanun, V. et al. Development of antibodies to Burkholderia
pseudomallei during childhood in melioidosis-endemic northeast Thailand.
Am. J. Trop. Med. Hyg 74, 1074-1075 (2006).
244. Khupulsup, K. & Petchclai, B. Application of indirect hemagglutination test
and indirect fluorescent antibody test for IgM antibody for diagnosis of
melioidosis in Thailand. Am. J. Trop. Med. Hyg 35, 366-369 (1986).
245. Yap, E.H. et al. Serodiagnosis of melioidosis in Singapore by the indirect
haemagglutination test. Singapore Med J 32, 211-213 (1991).
246. Charoenwong, P., Lumbiganon, P. & Puapermpoonsiri, S. The prevalence
of the indirect hemagglutination test for melioidosis in children in an
endemic area. Southeast Asian J. Trop. Med. Public Health 23, 698-701
(1992).
247. Norazah, A., Rohani, M.Y., Chang, P.T. & Kamel, A.G. Indirect
hemagglutination antibodies against Burkholderia pseudomallei in normal
170

lood donors and suspected cases of melioidosis in Malaysia. Southeast
Asian J. Trop. Med. Public Health 27, 263-266 (1996).
248. O’Brien, M. et al. Further evaluation of a rapid diagnostic test for
melioidosis in an area of endemicity. J. Clin. Microbiol 42, 2239-2240
(2004).
249. Warner, J.M., Pelowa, D.B., Currie, B.J. & Hirst, R.G. Melioidosis in a rural
community of Western Province, Papua New Guinea. Trans. R. Soc. Trop.
Med. Hyg 101, 809-813 (2007).
250. Lazzaroni, S.M. et al. Seropositivity to Burkholderia pseudomallei does not
reflect the development of cell-mediated immunity. Trans. R. Soc. Trop.
Med. Hyg 102 Suppl 1, S66-70 (2008).
251. Wuthiekanun, V. et al. Burkholderia pseudomallei antibodies in children,
Cambodia. Emerging Infect. Dis 14, 301-303 (2008).
252. Tiyawisutsri, R. et al. Antibodies from patients with melioidosis recognize
Burkholderia mallei but not Burkholderia thailandensis antigens in the
indirect hemagglutination assay. J. Clin. Microbiol 43, 4872-4874 (2005).
253. Gilmore, G., Barnes, J., Ketheesan, N. & Norton, R. Use of antigens derived
from Burkholderia pseudomallei, B. thailandensis, and B. cepacia in the
indirect hemagglutination assay for melioidosis. Clin. Vaccine Immunol 14,
1529-1531 (2007).
254. Dance, D.A. et al. The use of bone marrow culture for the diagnosis of
melioidosis. Trans. R. Soc. Trop. Med. Hyg 84, 585-587 (1990).
255. Cravitz, L. & Miller, W.R. Immunologic studies with Malleomyces mallei
and Malleomyces pseudomallei, agglutination and complement fixation
tests in man and laboratory animals. J. Infect. Dis 86, 52-62 (1950).
256. Alexander, A.D., Huxsoll, D.L., Warner, A.R., Shepler, V. & Dorsey, A.
Serological diagnosis of human melioidosis with indirect hemagglutination
and complement fixation tests. Appl Microbiol 20, 825-833 (1970).
257. Peacock, S.J., Cheng, A.C., Currie, B.J. & Dance, D.A.B. The use of positive
serological tests as evidence of exposure to Burkholderia pseudomallei.
Am. J. Trop. Med. Hyg 84, 1021-1022; author reply 1023 (2011).
258. Nigg, C. Serological studies on subclinical melioidosis. J. Immunol. 91, 18–
28 (1963).
171

259. White, N.J. et al. Halving of mortality of severe melioidosis by ceftazidime.
Lancet 2, 697-701 (1989).
260. Teparrugkul, P. Indirect hemagglutination test for the diagnosis of
melioidosis in Ubon Ratchathani. J. Infect. Dis. Antimicrob. Agents 14, 15–
17 (1997).
261. Cheng, A.C., O’brien, M., Freeman, K., Lum, G. & Currie, B.J. Indirect
hemagglutination assay in patients with melioidosis in northern Australia.
Am. J. Trop. Med. Hyg 74, 330-334 (2006).
262. Limmathurotsakul, D. et al. Enzyme-linked immunosorbent assay for the
diagnosis of melioidosis: better than we thought. Clin. Infect. Dis 52, 1024-
1028 (2011).
263. Appassakij, H., Silpapojakul, K.R., Wansit, R. & Pornpatkul, M. Diagnostic
value of the indirect hemagglutination test for melioidosis in an endemic
area. Am. J. Trop. Med. Hyg 42, 248-253 (1990).
264. Cheng, A.C. et al. Prospective evaluation of a rapid immunochromogenic
cassette test for the diagnosis of melioidosis in northeast Thailand. Trans.
R. Soc. Trop. Med. Hyg 100, 64-67 (2006).
265. Puapermpoonsiri, S., Puapermpoonsiri, P., Bhuripanyo, K., Vilachai, C. &
Auncharoen, A. Indirect hemagglutination antibody titer to Pseudomonas
pseudomallei in patients with melioidosis. Melioidosis 193–6 (1986).
266. UNICEF At a glance: Singapore. at
267. Puthucheary, S.D., Anuar, A.S.S. & Tee, T.S. Burkholderia thailandensis
whole cell antigen cross-reacts with B. pseudomallei antibodies from
patients with melioidosis in an immunofluorescent assay. Southeast Asian
J. Trop. Med. Public Health 41, 395-400 (2010).
268. Thepthai, C., Smithtikarn, S., Suksuwan, M., Songsivilai, S. & Dharakul, T.
Serodiagnosis of melioidosis by a competitive enzyme-linked
immunosorbent assay using a lipopolysaccharide-specific monoclonal
antibody. Asian Pac. J. Allergy Immunol 23, 127-132 (2005).
269. Phung, L.V. et al. Enzyme-linked immunosorbent assay (ELISA) using a
glycolipid antigen for the serodiagnosis of melioidosis. FEMS Immunol.
Med. Microbiol 12, 259-264 (1995).
172

270. Wuthiekanun, V. et al. Evaluation of immunoglobulin M (IgM) and IgG
rapid cassette test kits for diagnosis of melioidosis in an area of endemicity.
J. Clin. Microbiol 42, 3435-3437 (2004).
271. Chuah, S.C., Gilmore, G. & Norton, R.E. Rapid serological diagnosis of
melioidosis: an evaluation of a prototype immunochromatographic test.
Pathology 37, 169-171 (2005).
272. Splettstoesser, W.D., Frangoulidis, D. & Puthucheary, S.D. Validation and
comparison of an extrapolysaccharide (EPS)-based in-house ELISA and the
PanBio melioidosis rapid cassette test-kits for serodiagnosis of melioidosis
in a non-endemic area. Trans. R. Soc. Trop. Med. Hyg 102 Suppl 1, S45-
46 (2008).
273. Limmathurotsakul, D. et al. Defining the true sensitivity of culture for the
diagnosis of melioidosis using Bayesian latent class models. PLoS ONE 5,
e12485 (2010).
274. Peck, C.R. & Zwanenburg, T. Melioidosis presenting as abscess in neck. Br
Med J 1, 337–338 (1947).
275. Miller, W.R., Pannell, L. & Ingalls, M.S. Experimental chemotherapy in
glanders and melioidosis. Am J Hyg 47, 205-213 (1948).
276. Mirick, G.S., Zimmerman, H.S., Maner, G.D. & Humphrey, A. Melioidosis
on Guam. JAMA 130, 1063–1067 (1946).
277. McDowell, F. & Varney, P.L. Melioidosis, report of first case from the
Western Hemisphere. J Am Med Assoc 134, 361 (1947).
278. Cros, R. & Demarchi, J. Note sur l’action in vitro de la chloromycetine sur
le bacille de Whitmore. Ann Inst Pasteur (Paris) 79, 217-221 (1950).
279. Hall, W.H. & Manion, R.E. Antibiotic susceptibility of Pseudomonas
pseudomallei. Antimicrob. Agents Chemother 4, 193-195 (1973).
280. Brygoo, E.R. & Henry, E. Action in vitro de divers antibiotiques sur le
bacille de Whitmore. Bull Soc Pathol Exot Filiales 46, 279-283 (1953).
281. Chambon, L., De Lajudie, P. & Fournier, J. Etude de la sensibilité du bacille
de Whitmore aux antibiotiques in vitro et chez les malades atteints de
mélioidose. Bull Soc Pathol Exot Filiales 47, 139-153 (1954).
282. Dunlop, S.J.C. Rapid recovery in a case of melioidosis. Documenta de
medicina geographica et tropica 4, 296–300 (1952).
173

283. Chambon, L. Resistance du bacille de Whitmore acquise in vitro et in vivo a
l’egard du chloramphenicol, de l’aureomycine et de la terramycine. Ann
Inst Pasteur (Paris) 88, 315-324 (1955).
284. Tan, A.L. & Tan, M.-L. Melioidosis: antibiogram of cases in Singapore
1987-2007. Trans. R. Soc. Trop. Med. Hyg 102 Suppl 1, S101-102 (2008).
285. Moore, R.A., DeShazer, D., Reckseidler, S., Weissman, A. & Woods, D.E.
Efflux-mediated aminoglycoside and macrolide resistance in Burkholderia
pseudomallei. Antimicrob. Agents Chemother 43, 465-470 (1999).
286. Burtnick, M.N. & Woods, D.E. Isolation of polymyxin B-susceptible
mutants of Burkholderia pseudomallei and molecular characterization of
genetic loci involved in polymyxin B resistance. Antimicrob. Agents
Chemother 43, 2648-2656 (1999).
287. Dance, D.A., Wuthiekanun, V., Chaowagul, W. & White, N.J. The
antimicrobial susceptibility of Pseudomonas pseudomallei. Emergence of
resistance in vitro and during treatment. J. Antimicrob. Chemother 24,
295-309 (1989).
288. Cheng, A.C., West, T.E. & Peacock, S.J. Surviving sepsis in developing
countries. Crit. Care Med 36, 2487 (2008).
289. Kollef, M.H., Sherman, G., Ward, S. & Fraser, V.J. Inadequate
antimicrobial treatment of infections: a risk factor for hospital mortality
among critically ill patients. Chest 115, 462-474 (1999).
290. Ibrahim, E.H., Sherman, G., Ward, S., Fraser, V.J. & Kollef, M.H. The
influence of inadequate antimicrobial treatment of bloodstream infections
on patient outcomes in the ICU setting. Chest 118, 146-155 (2000).
291. Smith, M.D., Wuthiekanun, V., Walsh, A.L. & White, N.J. In-vitro activity
of carbapenem antibiotics against beta-lactam susceptible and resistant
strains of Burkholderia pseudomallei. J. Antimicrob. Chemother 37, 611-
615 (1996).
292. Simpson, A.J. et al. Differential antibiotic-induced endotoxin release in
severe melioidosis. J. Infect. Dis 181, 1014-1019 (2000).
293. Chierakul, W. et al. Two randomized controlled trials of ceftazidime alone
versus ceftazidime in combination with trimethoprim-sulfamethoxazole for
the treatment of severe melioidosis. Clin. Infect. Dis 41, 1105-1113 (2005).
174

294. Chierakul, W. et al. Addition of trimethoprim-sulfamethoxazole to
ceftazidime during parenteral treatment of melioidosis is not associated
with a long-term outcome benefit. Clin. Infect. Dis 45, 521-523 (2007).
295. Ulett, G.C., Hirst, R., Bowden, B., Powell, K. & Norton, R. A comparison of
antibiotic regimens in the treatment of acute melioidosis in a mouse model.
J. Antimicrob. Chemother 51, 77-81 (2003).
296. Suputtamongkol, Y. et al. Ceftazidime vs. amoxicillin/clavulanate in the
treatment of severe melioidosis. Clin. Infect. Dis 19, 846-853 (1994).
297. Jenney, A.W., Lum, G., Fisher, D.A. & Currie, B.J. Antibiotic susceptibility
of Burkholderia pseudomallei from tropical northern Australia and
implications for therapy of melioidosis. Int. J. Antimicrob. Agents 17, 109-
113 (2001).
298. Dance, D.A., Wuthiekanun, V., Chaowagul, W., Suputtamongkol, Y. &
White, N.J. Development of resistance to ceftazidime and co-amoxiclav in
Pseudomonas pseudomallei. J. Antimicrob. Chemother 28, 321-324 (1991).
299. Thamlikitkul, V. & Trakulsomboon, S. In vitro activity of doripenem
against Burkholderia pseudomallei. Antimicrob Agents Chemother
(2009).doi:10.1128/AAC.00893-08
300. Shih, H.-I. et al. Sporadic and outbreak cases of melioidosis in southern
Taiwan: clinical features and antimicrobial susceptibility. Infection 37, 9-15
(2009).
301. Thamlikitkul, V. & Trakulsomboon, S. In vitro activity of tigecycline against
Burkholderia pseudomallei and Burkholderia thailandensis. Antimicrob
Agents Chemother 50, 1555-7 (2006).
302. Feterl, M., Govan, B., Engler, C., Norton, R. & Ketheesan, N. Activity of
tigecycline in the treatment of acute Burkholderia pseudomallei infection in
a murine model. Int J Antimicrob Agents 28, 460-4 (2006).
303. Sam, I.-C., See, K.H. & Puthucheary, S.D. Susceptibility of Burkholderia
pseudomallei to tigecycline and other antimicrobials. Diagn. Microbiol.
Infect. Dis 67, 308-309 (2010).
304. Sivalingam, S.P., Sim, S.H., Aw, L.T. & Ooi, E.E. Antibiotic susceptibility of
50 clinical isolates of Burkholderia pseudomallei from Singapore. J.
Antimicrob. Chemother 58, 1102-1103 (2006).
175

305. Karunakaran, R. & Puthucheary, S.D. Burkholderia pseudomallei: In vitro
susceptibility to some new and old antimicrobials. Scand J Infect Dis 39,
858-861 (2007).
306. Limmathurotsakul, D. et al. Risk factors for recurrent melioidosis in
northeast Thailand. Clin. Infect. Dis 43, 979-986 (2006).
307. Chaowagul, W., Suputtamongkul, Y., Smith, M.D. & White, N.J. Oral
fluoroquinolones for maintenance treatment of melioidosis. Trans. R. Soc.
Trop. Med. Hyg 91, 599-601 (1997).
308. Chaowagul, W. et al. A comparison of chloramphenicol, trimethoprim-
sulfamethoxazole, and doxycycline with doxycycline alone as maintenance
therapy for melioidosis. Clin. Infect. Dis 29, 375-380 (1999).
309. Chetchotisakd, P., Chaowagul, W., Mootsikapun, P., Budhsarawong, D. &
Thinkamrop, B. Maintenance therapy of melioidosis with ciprofloxacin plus
azithromycin compared with cotrimoxazole plus doxycycline. Am. J. Trop.
Med. Hyg 64, 24-27 (2001).
310. Chaowagul, W. et al. Open-label randomized trial of oral trimethoprim-
sulfamethoxazole, doxycycline, and chloramphenicol compared with
trimethoprim-sulfamethoxazole and doxycycline for maintenance therapy
of melioidosis. Antimicrob. Agents Chemother 49, 4020-4025 (2005).
311. Berrie, C. Cotrimoxazole as effective without doxycycline for eradication of
melioidosis. DocGuide.com (2011).at
312. Dance, D.A., Wuthiekanun, V., Chaowagul, W. & White, N.J. Interactions
in vitro between agents used to treat melioidosis. J. Antimicrob.
Chemother 24, 311-316 (1989).
313. Cheng, A.C. et al. Dosing regimens of cotrimoxazole (trimethoprim-
sulfamethoxazole) for melioidosis. Antimicrob. Agents Chemother
(2009).doi:10.1128/AAC.01301-08
314. Sarkar-Tyson, M. & Titball, R.W. Progress toward development of vaccines
against melioidosis: A review. Clin Ther 32, 1437-1445 (2010).
315. Patel, N. et al. Development of vaccines against Burkholderia
pseudomallei. Front Microbiol 2, 198 (2011).
176

316. Peacock, S.J. et al. Management of accidental laboratory exposure to
Burkholderia pseudomallei and B. mallei. Emerging Infect. Dis 14, e2
(2008).
317. Wuthiekanun, V. et al. Perasafe, Virkon and bleach are bactericidal for
Burkholderia pseudomallei, a select agent and the cause of melioidosis. J.
Hosp. Infect 77, 183-184 (2011).
318. Gal, D. et al. Contamination of hand wash detergent linked to
occupationally acquired melioidosis. Am. J. Trop. Med. Hyg 71, 360-362
(2004).
319. Sookpranee, M., Lumbiganon, P., Puapermpoonsiri, S., Tattawasatra, A. &
Nopwinyoovongs, J. Contamination of Savlon solution with Pseudomonas
pseudomallei at Srinagarind Hospital. Melioidosis 211–213 (1989).
320. Howard, K. & Inglis, T.J.J. The effect of free chlorine on Burkholderia
pseudomallei in potable water. Water Res 37, 4425-4432 (2003).
321. Inglis, T.J.J., O’Reilly, L., Merritt, A.J., Levy, A. & Heath, C. The aftermath
of the Western Australian melioidosis outbreak. Am. J. Trop. Med. Hyg
84, 851-857 (2011).
322. O’Connell, H.A. et al. Variability of Burkholderia pseudomallei strain
sensitivities to chlorine disinfection. Appl Environ Microbiol 75, 5405-
5409 (2009).
323. Cheng, A.C. et al. A randomized controlled trial of granulocyte colony-
stimulating factor for the treatment of severe sepsis due to melioidosis in
Thailand. Clin. Infect. Dis 45, 308-314 (2007).
324. Metcalf, D. The granulocyte-macrophage colony-stimulating factors.
Science 229, 16-22 (1985).
325. Easton, A., Haque, A., Chu, K., Lukaszewski, R. & Bancroft, G.J. A critical
role for neutrophils in resistance to experimental infection with
Burkholderia pseudomallei. J. Infect. Dis 195, 99-107 (2007).
326. Cheng, A.C., Stephens, D.P., Anstey, N.M. & Currie, B.J. Adjunctive
granulocyte colony-stimulating factor for treatment of septic shock due to
melioidosis. Clin. Infect. Dis 38, 32-37 (2004).
327. Finney, S.J., Zekveld, C., Elia, A. & Evans, T.W. Glucose control and
mortality in critically ill patients. JAMA 290, 2041-2047 (2003).
177

328. Koh, G.C.K.W. et al. Glyburide is anti-inflammatory and associated with
reduced mortality in melioidosis. Clin Infect Dis 52, 717–725 (2011).
329. Bellomo, R. & Egi, M. Glycemic control in the intensive care unit: why we
should wait for NICE-SUGAR. Mayo Clin. Proc 80, 1546-1548 (2005).
330. van den Berghe, G. et al. Intensive insulin therapy in the medical ICU. N.
Engl. J. Med 354, 449-461 (2006).
331. Malmberg, K. et al. Intense metabolic control by means of insulin in
patients with diabetes mellitus and acute myocardial infarction (DIGAMI
2): effects on mortality and morbidity. Eur. Heart J 26, 650-661 (2005).
332. Collier, B. et al. The impact of a normoglycemic management protocol on
clinical outcomes in the trauma intensive care unit. JPEN J Parenter
Enteral Nutr 29, 353-358; discussion 359 (2005).
333. Finfer, S. et al. Intensive versus conventional glucose control in critically ill
patients. N. Engl. J. Med 360, 1283-1297 (2009).
334. Griesdale, D.E.G. et al. Intensive insulin therapy and mortality among
critically ill patients: a meta-analysis including NICE-SUGAR study data.
CMAJ 180, 821-827 (2009).
335. Warren, H.S. Strategies for the treatment of sepsis. N. Engl. J. Med 336,
952-953 (1997).
336. Thomas, L. Notes of a Biology Watcher: Germs. N. Engl. J. Med 287, 553-
555 (1972).
337. Hinshaw, L.B. et al. Survival of primates in LD100 septic shock following
therapy with antibody to tumor necrosis factor (TNF alpha). Circ. Shock
30, 279-292 (1990).
338. Hinshaw, L.B. et al. Lethal Staphylococcus aureus-induced shock in
primates: prevention of death with anti-TNF antibody. J Trauma 33, 568-
573 (1992).
339. Ohlsson, K., Björk, P., Bergenfeldt, M., Hageman, R. & Thompson, R.C.
Interleukin-1 receptor antagonist reduces mortality from endotoxin shock.
Nature 348, 550-552 (1990).
340. Alexander, H.R., Doherty, G.M., Buresh, C.M., Venzon, D.J. & Norton, J.A.
A recombinant human receptor antagonist to interleukin 1 improves
survival after lethal endotoxemia in mice. J. Exp. Med 173, 1029-1032
(1991).
178

341. Wakabayashi, G., Gelfand, J.A., Burke, J.F., Thompson, R.C. & Dinarello,
C.A. A specific receptor antagonist for interleukin 1 prevents Escherichia
coli-induced shock in rabbits. FASEB J 5, 338-343 (1991).
342. Fischer, E. et al. Interleukin-1 receptor blockade improves survival and
hemodynamic performance in Escherichia coli septic shock, but fails to
alter host responses to sublethal endotoxemia. J. Clin. Invest 89, 1551-1557
(1992).
343. Alexander, H.R. et al. Recombinant interleukin-1 receptor antagonist (IL-
1ra): effective therapy against gram-negative sepsis in rats. Surgery 112,
188-193; discussion 193-194 (1992).
344. Sprung, C.L. et al. The effects of high-dose corticosteroids in patients with
septic shock. A prospective, controlled study. N. Engl. J. Med 311, 1137-
1143 (1984).
345. Bone, R.C., Fisher, C.J., Clemmer, T.P., Slotman, G.J. & Metz, C.A. Early
methylprednisolone treatment for septic syndrome and the adult
respiratory distress syndrome. Chest 92, 1032-1036 (1987).
346. Luce, J.M. et al. Ineffectiveness of high-dose methylprednisolone in
preventing parenchymal lung injury and improving mortality in patients
with septic shock. Am. Rev. Respir. Dis 138, 62-68 (1988).
347. Slotman, G.J., Fisher, C.J., Bone, R.C., Clemmer, T.P. & Metz, C.A.
Detrimental effects of high-dose methylprednisolone sodium succinate on
serum concentrations of hepatic and renal function indicators in severe
sepsis and septic shock. The Methylprednisolone Severe Sepsis Study
Group. Crit. Care Med 21, 191-195 (1993).
348. Annane, D. et al. Effect of treatment with low doses of hydrocortisone and
fludrocortisone on mortality in patients with septic shock. JAMA 288,
862-871 (2002).
349. Briegel, J. et al. Stress doses of hydrocortisone reverse hyperdynamic septic
shock: a prospective, randomized, double-blind, single-center study. Crit.
Care Med 27, 723-732 (1999).
350. Sprung, C.L. et al. Hydrocortisone therapy for patients with septic shock.
N. Engl. J. Med 358, 111-124 (2008).
351. Arabi, Y.M. et al. Low-dose hydrocortisone in patients with cirrhosis and
septic shock: a randomized controlled trial. CMAJ 182, 1971-1977 (2010).
179

352. Tracey, K.J. et al. Anti-cachectin/TNF monoclonal antibodies prevent
septic shock during lethal bacteraemia. Nature 330, 662-664 (1987).
353. Fong, Y. et al. Antibodies to cachectin/tumor necrosis factor reduce
interleukin 1 beta and interleukin 6 appearance during lethal bacteremia. J.
Exp. Med 170, 1627-1633 (1989).
354. Abraham, E. et al. Efficacy and safety of monoclonal antibody to human
tumor necrosis factor alpha in patients with sepsis syndrome. A
randomized, controlled, double-blind, multicenter clinical trial. TNF-alpha
MAb Sepsis Study Group. JAMA 273, 934-941 (1995).
355. Abraham, E. et al. p55 Tumor necrosis factor receptor fusion protein in the
treatment of patients with severe sepsis and septic shock. A randomized
controlled multicenter trial. Ro 45-2081 Study Group. JAMA 277, 1531-
1538 (1997).
356. Abraham, E. et al. Double-blind randomised controlled trial of monoclonal
antibody to human tumour necrosis factor in treatment of septic shock.
NORASEPT II Study Group. Lancet 351, 929-933 (1998).
357. Abraham, E. et al. Lenercept (p55 tumor necrosis factor receptor fusion
protein) in severe sepsis and early septic shock: a randomized, double-
blind, placebo-controlled, multicenter phase III trial with 1,342 patients.
Crit. Care Med 29, 503-510 (2001).
358. Cohen, J. & Carlet, J. INTERSEPT: an international, multicenter, placebo-
controlled trial of monoclonal antibody to human tumor necrosis factor-
alpha in patients with sepsis. International Sepsis Trial Study Group. Crit.
Care Med 24, 1431-1440 (1996).
359. Lertmemongkolchai, G., Cai, G., Hunter, C.A. & Bancroft, G.J. Bystander
activation of CD8+ T cells contributes to the rapid production of IFN-
gamma in response to bacterial pathogens. J. Immunol 166, 1097-1105
(2001).
360. Fisher, C.J. et al. Initial evaluation of human recombinant interleukin-1
receptor antagonist in the treatment of sepsis syndrome: a randomized,
open-label, placebo-controlled multicenter trial. Crit. Care Med 22, 12-21
(1994).
361. Fisher, C.J. et al. Recombinant human interleukin 1 receptor antagonist in
the treatment of patients with sepsis syndrome. Results from a
180

andomized, double-blind, placebo-controlled trial. Phase III rhIL-1ra
Sepsis Syndrome Study Group. JAMA 271, 1836-1843 (1994).
362. Opal, S.M. et al. Confirmatory interleukin-1 receptor antagonist trial in
severe sepsis: a phase III, randomized, double-blind, placebo-controlled,
multicenter trial. The Interleukin-1 Receptor Antagonist Sepsis
Investigator Group. Crit. Care Med 25, 1115-1124 (1997).
363. Masiello, S.A. & Siegel, J.P. License 125029/0. (2001).at
364. Hosac, A.M. Drotrecogin alfa (activated): the first FDA-approved treatment
for severe sepsis. Proc (Bayl Univ Med Cent) 15, 224-227 (2002).
365. Bernard, G.R. et al. Efficacy and safety of recombinant human activated
protein C for severe sepsis. N. Engl. J. Med 344, 699-709 (2001).
366. LaRosa, S.P. et al. Decreased protein C, protein S, and antithrombin levels
are predictive of poor outcome in Gram-negative sepsis caused by
Burkholderia pseudomallei. Int. J. Infect. Dis 10, 25-31 (2006).
367. Wiersinga, W.J. et al. Activation of coagulation with concurrent
impairment of anticoagulant mechanisms correlates with a poor outcome
in severe melioidosis. J. Thromb. Haemost 6, 32-39 (2008).
368. Brown, A.E. et al. Immune cell activation in melioidosis: increased serum
levels of interferon-gamma and soluble interleukin-2 receptors without
change in soluble CD8 protein. J. Infect. Dis 163, 1145-1148 (1991).
369. Propst, K.L., Troyer, R.M., Kellihan, L.M., Schweizer, H.P. & Dow, S.W.
Immunotherapy markedly increases the effectiveness of antimicrobial
therapy for treatment of Burkholderia pseudomallei infection. Antimicrob
Agents Chemother (2010).doi:10.1128/AAC.01513-09
370. Gao, X.-F., Yang, Z.-W. & Li, J. Adjunctive therapy with interferon-gamma
for the treatment of pulmonary tuberculosis: a systematic review. Int J
Infect Dis (2011).doi:10.1016/j.ijid.2011.05.002
371. Koh, G.C.K.W. & Limmathurotsakul, D. Gamma interferon
supplementation for melioidosis. Antimicrob. Agents Chemother 54, 4520
(2010).
372. Santanirand, P., Harley, V.S., Dance, D.A., Drasar, B.S. & Bancroft, G.J.
Obligatory role of gamma interferon for host survival in a murine model of
181

infection with Burkholderia pseudomallei. Infect. Immun 67, 3593-3600
(1999).
373. Greenawald, K.A., Nash, G. & Foley, F.D. Acute systemic melioidosis.
Autopsy findings in four patients. Am. J. Clin. Pathol 52, 188-198 (1969).
374. Piggott, J.A. & Hochholzer, L. Human melioidosis. A histopathologic study
of acute and chronic melioidosis. Arch Pathol 90, 101-111 (1970).
375. Lauw, F.N. et al. Elevated plasma concentrations of interferon (IFN)-
gamma and the IFN-gamma-inducing cytokines interleukin (IL)-18, IL-12,
and IL-15 in severe melioidosis. J. Infect. Dis 180, 1878-1885 (1999).
376. Lauw, F.N. et al. The CXC chemokines gamma interferon (IFN-gamma)-
inducible protein 10 and monokine induced by IFN-gamma are released
during severe melioidosis. Infect. Immun 68, 3888-3893 (2000).
377. Calandra, T. et al. Prognostic values of tumor necrosis factor/cachectin,
interleukin-1, interferon-alpha, and interferon-gamma in the serum of
patients with septic shock. Swiss-Dutch J5 Immunoglobulin Study Group.
J. Infect. Dis 161, 982-987 (1990).
378. Pinsky, M.R. et al. Serum cytokine levels in human septic shock. Relation
to multiple-system organ failure and mortality. Chest 103, 565-575 (1993).
379. Ahmed, K. et al. Increased serum levels of interferon-gamma and
interleukin-12 during human brucellosis. Am. J. Trop. Med. Hyg 61, 425-
427 (1999).
380. Brown, A.E., Rieder, K.T. & Webster, H.K. Prolonged elevations of soluble
interleukin-2 receptors in tuberculosis. Am. Rev. Respir. Dis 139, 1036-
1038 (1989).
381. He, X.-Y. et al. T regulatory cells and Th1/Th2 cytokines in peripheral
blood from tuberculosis patients. Eur. J. Clin. Microbiol. Infect. Dis 29,
643-650 (2010).
382. Kornelisse, R.F. et al. Intrathecal production of interleukin-12 and gamma
interferon in patients with bacterial meningitis. Infect. Immun 65, 877-881
(1997).
383. Bjerre, A., Brusletto, B., Høiby, E.A., Kierulf, P. & Brandtzaeg, P. Plasma
interferon-gamma and interleukin-10 concentrations in systemic
meningococcal disease compared with severe systemic Gram-positive
septic shock. Crit. Care Med 32, 433-438 (2004).
182

384. Hazelzet, J.A. et al. Interleukin 12 levels during the initial phase of septic
shock with purpura in children: relation to severity of disease. Cytokine 9,
711-716 (1997).
385. Sheikh, A. et al. Interferon-γ and proliferation responses to Salmonella
enterica serotype Typhi proteins in patients with S. Typhi bacteremia in
Dhaka, Bangladesh. PLoS Negl Trop Dis 5, e1193 (2011).
386. Murray, H.W., Rubin, B.Y., Carriero, S.M., Harris, A.M. & Jaffee, E.A.
Human mononuclear phagocyte antiprotozoal mechanisms: oxygen-
dependent vs oxygen-independent activity against intracellular
Toxoplasma gondii. J. Immunol 134, 1982-1988 (1985).
387. Nacy, C.A., Fortier, A.H., Meltzer, M.S., Buchmeier, N.A. & Schreiber, R.D.
Macrophage activation to kill Leishmania major: activation of macrophages
for intracellular destruction of amastigotes can be induced by both
recombinant interferon-gamma and non-interferon lymphokines. J.
Immunol 135, 3505-3511 (1985).
388. Buchmeier, N.A. & Schreiber, R.D. Requirement of endogenous interferon-
gamma production for resolution of Listeria monocytogenes infection.
Proc. Natl. Acad. Sci. U.S.A 82, 7404-7408 (1985).
389. Henle, W. & Henle, G. The effect of ultraviolet irradiation on various
properties. J. Exp. Med. 85, 347–364 (1947).
390. Isaacs, A. & Lindenmann, J. Virus interference. I. The Interferon.
Proceedings of the Royal Society B 147, 258–267 (1957).
391. Havell, E.A. et al. Two antigenically distinct species of human interferon.
Proc. Natl. Acad. Sci. U.S.A 72, 2185-2187 (1975).
392. Stewart II, W.E. et al. Distinct molecular species of human interferons:
requirements for stabilzation and reactivation of human leukocyte and
fibroblast interferons. J. Gen. Virol 26, 327-331 (1975).
393. Grob, P.M. & Chadha, K.C. Separation of human leukocyte interferon
components by concanavalin A-agarose affinity chromatography and their
characterization. Biochemistry 18, 5782-5786 (1979).
394. Nathan, C.F., Remold, H.G. & David, J.R. Characterization of a lymphocyte
factor which alters macrophage functions. J. Exp. Med 137, 275-290
(1973).
183

395. Gray, P.W. & Goeddel, D.V. Structure of the human immune interferon
gene. Nature 298, 859-863 (1982).
396. Gray, P.W. & Goeddel, D.V. Human immune interferon (IFN-gamma) gene
sequence and structure. Basic Life Sci 25, 35-61 (1983).
397. Kleinschmidt, W.J. & Schultz, R.M. Similarities of murine gamma
interferon and the lymphokine that renders macrophages cytotoxic. J.
Interferon Res 2, 291-299 (1982).
398. Roberts, W.K. & Vasil, A. Evidence for the identity of murine gamma
interferon and macrophage activating factor. J. Interferon Res 2, 519-532
(1982).
399. Schultz, R.M. & Kleinschmidt, W.J. Functional identity between murine
gamma interferon and macrophage activating factor. Nature 305, 239-240
(1983).
400. Fukazawa, Y. et al. Biological and biochemical characterization of
macrophage activating factor (MAF) in murine lymphocytes:
physiocochemical similarity of MAF to gamma interferon (IFN-gamma).
Microbiol. Immunol 28, 691-702 (1984).
401. Svedersky, L.P. et al. Biological and antigenic similarities of murine
interferon-gamma and macrophage-activating factor. J. Exp. Med 159,
812-827 (1984).
402. Wheeler, A.P. & Bernard, G.R. Treating patients with severe sepsis. N.
Engl. J. Med 340, 207-214 (1999).
403. Heinzel, F.P. The role of IFN-gamma in the pathology of experimental
endotoxemia. J. Immunol 145, 2920-2924 (1990).
404. Car, B.D. et al. Interferon gamma receptor deficient mice are resistant to
endotoxic shock. J. Exp. Med 179, 1437-1444 (1994).
405. Fink, M.P. & Heard, S.O. Laboratory models of sepsis and septic shock. J.
Surg. Res 49, 186-196 (1990).
406. Poli-de-Figueiredo, L.F., Garrido, A.G., Nakagawa, N. & Sannomiya, P.
Experimental models of sepsis and their clinical relevance. Shock 30
Suppl 1, 53-59 (2008).
407. Zeni, F., Freeman, B. & Natanson, C. Anti-inflammatory therapies to treat
sepsis and septic shock: a reassessment. Crit. Care Med 25, 1095-1100
(1997).
184

408. Belikoff, B. & Buras, J.A. A practical approach to animal models of sepsis.
Sourcebook of Models for Biomedical Research 473–482 (2008).at
409. Koo, G.C. & Gan, Y.-H. The innate interferon gamma response of BALB/c
and C57BL/6 mice to in vitro Burkholderia pseudomallei infection. BMC
Immunol 7, 19 (2006).
410. Haque, A. et al. Role of T cells in innate and adaptive immunity against
murine Burkholderia pseudomallei infection. J. Infect. Dis 193, 370-379
(2006).
411. Wiersinga, W.J. et al. Endogenous interleukin-18 improves the early
antimicrobial host response in severe melioidosis. Infect. Immun 75, 3739-
3746 (2007).
412. Friedland, J.S. et al. Prolonged elevation of interleukin-8 and interleukin-6
concentrations in plasma and of leukocyte interleukin-8 mRNA levels
during septicemic and localized Pseudomonas pseudomallei infection.
Infect. Immun 60, 2402-2408 (1992).
413. Trinchieri, G. Cytokines acting on or secreted by macrophages during
intracellular infection (IL-10, IL-12, IFN-gamma). Curr. Opin. Immunol 9,
17-23 (1997).
414. Hemmi, H. et al. Small anti-viral compounds activate immune cells via the
TLR7 MyD88-dependent signaling pathway. Nat. Immunol 3, 196-200
(2002).
415. Nguyen, K.B. et al. Coordinated and distinct roles for IFN-alpha beta, IL-
12, and IL-15 regulation of NK cell responses to viral infection. J. Immunol
169, 4279-4287 (2002).
416. West, T.E., Ernst, R.K., Jansson-Hutson, M.J. & Skerrett, S.J. Activation of
Toll-like receptors by Burkholderia pseudomallei. BMC Immunol 9, 46
(2008).
417. Suputtamongkol, Y., Kwiatkowski, D., Dance, D.A., Chaowagul, W. &
White, N.J. Tumor necrosis factor in septicemic melioidosis. J. Infect. Dis
165, 561-564 (1992).
418. Collart, M.A., Belin, D., Vassalli, J.D., de Kossodo, S. & Vassalli, P. Gamma
interferon enhances macrophage transcription of the tumor necrosis
185

factor/cachectin, interleukin 1, and urokinase genes, which are controlled
by short-lived repressors. J. Exp. Med 164, 2113-2118 (1986).
419. Myers, M.J., Ghildyal, N. & Schook, L.B. Endotoxin and interferon-gamma
differentially regulate the transcriptional levels of proto-oncogenes and
cytokine genes during the differentiation of colony-stimulating factor type-
1-derived macrophages. Immunology 85, 318-324 (1995).
420. Farber, J.M. Mig and IP-10: CXC chemokines that target lymphocytes. J.
Leukoc. Biol 61, 246-257 (1997).
421. Simpson, A.J. et al. Prognostic value of cytokine concentrations (tumor
necrosis factor-alpha, interleukin-6, and interleukin-10) and clinical
parameters in severe melioidosis. J. Infect. Dis 181, 621-625 (2000).
422. Shi, Z. et al. A novel Toll-like receptor that recognizes vesicular stomatitis
virus. J. Biol. Chem 286, 4517-4524 (2011).
423. Hedayat, M., Netea, M.G. & Rezaei, N. Targeting of Toll-like receptors: a
decade of progress in combating infectious diseases. Lancet Infect Dis 11,
702–712 (2011).
424. Kawai, T. & Akira, S. The role of pattern-recognition receptors in innate
immunity: update on Toll-like receptors. Nat. Immunol 11, 373-384
(2010).
425. Leulier, F. & Lemaitre, B. Toll-like receptors—taking an evolutionary
approach. Nat. Rev. Genet 9, 165-178 (2008).
426. Casanova, J.-L., Abel, L. & Quintana-Murci, L. Human TLRs and IL-1Rs in
host defense: natural insights from evolutionary, epidemiological, and
clinical genetics. Annu. Rev. Immunol 29, 447-491 (2011).
427. Wiersinga, W.J., Wieland, C.W., Roelofs, J.J.T.H. & van der Poll, T. MyD88
dependent signaling contributes to protective host defense against
Burkholderia pseudomallei. PLoS ONE 3, e3494 (2008).
428. Chin, C.-Y., Monack, D.M. & Nathan, S. Genome wide transcriptome
profiling of a murine acute melioidosis model reveals new insights into how
Burkholderia pseudomallei overcomes host innate immunity. BMC
Genomics 11, 672 (2010).
429. Li, Q. & Verma, I.M. NF-kappaB regulation in the immune system. Nat.
Rev. Immunol 2, 725-734 (2002).
186

430. Karin, M., Cao, Y., Greten, F.R. & Li, Z.-W. NF-kappaB in cancer: from
innocent bystander to major culprit. Nat. Rev. Cancer 2, 301-310 (2002).
431. Takeda, K. & Akira, S. TLR signaling pathways. Semin. Immunol 16, 3-9
(2004).
432. Burstein, E. & Duckett, C.S. Dying for NF-kappaB? Control of cell death by
transcriptional regulation of the apoptotic machinery. Curr. Opin. Cell Biol
15, 732-737 (2003).
433. Piva, R., Belardo, G. & Santoro, M.G. NF-kappaB: a stress-regulated switch
for cell survival. Antioxid. Redox Signal 8, 478-486 (2006).
434. Sutmuller, R.P.M. et al. Toll-like receptor 2 controls expansion and
function of regulatory T cells. J. Clin. Invest 116, 485-494 (2006).
435. Fitzgerald, K.A. & Chen, Z.J. Sorting out Toll signals. Cell 125, 834-836
(2006).
436. Kawai, T. & Akira, S. Toll-like receptors and their crosstalk with other
innate receptors in infection and immunity. Immunity 34, 637-650 (2011).
437. Akira, S., Uematsu, S. & Takeuchi, O. Pathogen recognition and innate
immunity. Cell 124, 783-801 (2006).
438. Muzio, M. et al. Differential expression and regulation of toll-like receptors
(TLR) in human leukocytes: selective expression of TLR3 in dendritic cells.
J. Immunol 164, 5998-6004 (2000).
439. Hornung, V. et al. Quantitative expression of toll-like receptor 1-10 mRNA
in cellular subsets of human peripheral blood mononuclear cells and
sensitivity to CpG oligodeoxynucleotides. J. Immunol 168, 4531-4537
(2002).
440. Poltorak, A. et al. Defective LPS signaling in C3H/HeJ and C57BL/10ScCr
mice: mutations in Tlr4 gene. Science 282, 2085-2088 (1998).
441. Anuntagool, N. et al. Lipopolysaccharide heterogeneity among
Burkholderia pseudomallei from different geographic and clinical origins.
Am. J. Trop. Med. Hyg 74, 348-352 (2006).
442. Schromm, A.B. et al. Molecular genetic analysis of an endotoxin
nonresponder mutant cell line: a point mutation in a conserved region of
MD-2 abolishes endotoxin-induced signaling. J. Exp. Med 194, 79-88
(2001).
187

443. West, T.E. et al. Toll-like receptor 4 region genetic variants are associated
with susceptibility to melioidosis. Genes Immun
(2011).doi:10.1038/gene.2011.49
444. Morici, L.A., Heang, J., Tate, T., Didier, P.J. & Roy, C.J. Differential
susceptibility of inbred mouse strains to Burkholderia thailandensis aerosol
infection. Microb. Pathog (2009).doi:10.1016/j.micpath.2009.10.004
445. West, T.E., Hawn, T.R. & Skerrett, S.J. Toll-like receptor signaling in
airborne Burkholderia thailandensis infection. Infect. Immun 77, 5612-
5622 (2009).
446. Utaisincharoen, P. et al. Kinetic studies of the production of nitric oxide
(NO) and tumour necrosis factor-alpha (TNF-alpha) in macrophages
stimulated with Burkholderia pseudomallei endotoxin. Clin. Exp. Immunol
122, 324-329 (2000).
447. Dziarski, R. & Gupta, D. Role of MD-2 in TLR2- and TLR4-mediated
recognition of Gram-negative and Gram-positive bacteria and activation of
chemokine genes. J. Endotoxin Res 6, 401-405 (2000).
448. Bernheiden, M. et al. LBP, CD14, TLR4 and the murine innate immune
response to a peritoneal Salmonella infection. J. Endotoxin Res 7, 447-450
(2001).
449. Wang, X. et al. Toll-like receptor 4 mediates innate immune responses to
Haemophilus influenzae infection in mouse lung. J. Immunol 168, 810-815
(2002).
450. van Westerloo, D.J. et al. Toll-like receptor 4 deficiency and acute
pancreatitis act similarly in reducing host defense during murine
Escherichia coli peritonitis. Crit. Care Med 33, 1036-1043 (2005).
451. Knapp, S. et al. Differential roles of CD14 and toll-like receptors 4 and 2 in
murine Acinetobacter pneumonia. Am. J. Respir. Crit. Care Med 173, 122-
129 (2006).
452. Ting, J.P.-Y. et al. The NLR gene family: a standard nomenclature.
Immunity 28, 285-287 (2008).
453. Hibino, T. et al. The immune gene repertoire encoded in the purple sea
urchin genome. Dev. Biol 300, 349-365 (2006).
454. Inohara, N. & Nuñez, G. NODs: intracellular proteins involved in
inflammation and apoptosis. Nat. Rev. Immunol 3, 371-382 (2003).
188

455. Strober, W., Murray, P.J., Kitani, A. & Watanabe, T. Signalling pathways
and molecular interactions of NOD1 and NOD2. Nat. Rev. Immunol 6, 9-
20 (2006).
456. Le Bourhis, L., Benko, S. & Girardin, S.E. Nod1 and Nod2 in innate
immunity and human inflammatory disorders. Biochem. Soc. Trans 35,
1479-1484 (2007).
457. Pudla, M., Kananuruk, A., Limposuwan, K., Sirisinha, S. & Utaisincharoen,
P. Nucleotide-binding oligomerization domain-containing protein 2
regulates suppressor of cytokine signaling 3 expression in Burkholderia
pseudomallei- infected mouse macrophage cell line RAW 264.7. Innate
Immun (2010).doi:10.1177/1753425910385484
458. Martinon, F., Burns, K. & Tschopp, J. The inflammasome: a molecular
platform triggering activation of inflammatory caspases and processing of
proIL-beta. Mol. Cell 10, 417-426 (2002).
459. Franchi, L., Eigenbrod, T., Muñoz-Planillo, R. & Nuñez, G. The
inflammasome: a caspase-1-activation platform that regulates immune
responses and disease pathogenesis. Nat. Immunol 10, 241-247 (2009).
460. Faustin, B. et al. Reconstituted NALP1 inflammasome reveals two-step
mechanism of caspase-1 activation. Mol. Cell 25, 713-724 (2007).
461. Grenier, J.M. et al. Functional screening of five PYPAF family members
identifies PYPAF5 as a novel regulator of NF-kappaB and caspase-1. FEBS
Lett 530, 73-78 (2002).
462. Wang, L. et al. PYPAF7, a novel PYRIN-containing Apaf1-like protein that
regulates activation of NF-kappa B and caspase-1-dependent cytokine
processing. J. Biol. Chem 277, 29874-29880 (2002).
463. Dowds, T.A. et al. Regulation of cryopyrin/Pypaf1 signaling by pyrin, the
familial Mediterranean fever gene product. Biochem. Biophys. Res.
Commun 302, 575-580 (2003).
464. Mariathasan, S. et al. Differential activation of the inflammasome by
caspase-1 adaptors ASC and Ipaf. Nature 430, 213-218 (2004).
465. Miao, E.A. et al. Cytoplasmic flagellin activates caspase-1 and secretion of
interleukin 1beta via Ipaf. Nat. Immunol 7, 569-575 (2006).
189

466. Franchi, L. et al. Cytosolic flagellin requires Ipaf for activation of caspase-1
and interleukin 1beta in salmonella-infected macrophages. Nat. Immunol
7, 576-582 (2006).
467. Miao, E.A. et al. Innate immune detection of the type III secretion
apparatus through the NLRC4 inflammasome. Proc. Natl. Acad. Sci. U.S.A
107, 3076-3080 (2010).
468. Lightfield, K.L. et al. Critical function for Naip5 in inflammasome
activation by a conserved carboxy-terminal domain of flagellin. Nat.
Immunol 9, 1171-1178 (2008).
469. Franchi, L. et al. Critical role for Ipaf in Pseudomonas aeruginosa-induced
caspase-1 activation. Eur. J. Immunol 37, 3030-3039 (2007).
470. Sutterwala, F.S. et al. Immune recognition of Pseudomonas aeruginosa
mediated by the IPAF/NLRC4 inflammasome. J. Exp. Med 204, 3235-
3245 (2007).
471. Re, F., Sahoo, M., Ceballos-Olivera, I., Bina, J. & Miller, M. Critical role of
the NLRC4 inflammasome and IL-18 during mouse lung infection with
virulent Burkholderia pseudomallei. Decoding Innate Immunity A157
(2011).
472. Sun, G.W., Lu, J., Pervaiz, S., Cao, W.P. & Gan, Y.-H. Caspase-1 dependent
macrophage death induced by Burkholderia pseudomallei. Cell. Microbiol
7, 1447-1458 (2005).
473. Breitbach, K. et al. Caspase-1 mediates resistance in murine melioidosis.
Infect Immun (2009).doi:10.1128/IAI.01257-08
474. Black, R.A. et al. Generation of biologically active interleukin-1 beta by
proteolytic cleavage of the inactive precursor. J. Biol. Chem 263, 9437-
9442 (1988).
475. Black, R.A., Kronheim, S.R., Merriam, J.E., March, C.J. & Hopp, T.P. A
pre-aspartate-specific protease from human leukocytes that cleaves pro-
interleukin-1 beta. J. Biol. Chem 264, 5323-5326 (1989).
476. Kostura, M.J. et al. Identification of a monocyte specific pre-interleukin 1
beta convertase activity. Proc. Natl. Acad. Sci. U.S.A 86, 5227-5231 (1989).
477. Howard, A.D. et al. IL-1-converting enzyme requires aspartic acid residues
for processing of the IL-1 beta precursor at two distinct sites and does not
cleave 31-kDa IL-1 alpha. J. Immunol 147, 2964-2969 (1991).
190

478. Cerretti, D.P. et al. Molecular cloning of the interleukin-1 beta converting
enzyme. Science 256, 97-100 (1992).
479. Thornberry, N.A. et al. A novel heterodimeric cysteine protease is required
for interleukin-1 beta processing in monocytes. Nature 356, 768-774
(1992).
480. Thornberry, N.A. The caspase family of cysteine proteases. Br. Med. Bull
53, 478-490 (1997).
481. Alnemri, E.S. et al. Human ICE/CED-3 protease nomenclature. Cell 87, 171
(1996).
482. Saleh, M. et al. Enhanced bacterial clearance and sepsis resistance in
caspase-12-deficient mice. Nature 440, 1064-1068 (2006).
483. Xue, Y. et al. Spread of an inactive form of caspase-12 in humans is due to
recent positive selection. Am J Hum Genet 78, 659–670 (2006).
484. Martinon, F. & Tschopp, J. Inflammatory caspases: linking an intracellular
innate immune system to autoinflammatory diseases. Cell 117, 561-574
(2004).
485. Ghayur, T. et al. Caspase-1 processes IFN-gamma-inducing factor and
regulates LPS-induced IFN-gamma production. Nature 386, 619-623
(1997).
486. Gu, Y. et al. Activation of interferon-gamma inducing factor mediated by
interleukin-1beta converting enzyme. Science 275, 206-209 (1997).
487. Fantuzzi, G., Puren, A.J., Harding, M.W., Livingston, D.J. & Dinarello, C.A.
Interleukin-18 regulation of interferon gamma production and cell
proliferation as shown in interleukin-1beta-converting enzyme (caspase-1)-
deficient mice. Blood 91, 2118-2125 (1998).
488. Fantuzzi, G., Reed, D.A. & Dinarello, C.A. IL-12-induced IFN-gamma is
dependent on caspase-1 processing of the IL-18 precursor. J. Clin. Invest
104, 761-767 (1999).
489. Lamkanfi, M. et al. Targeted peptidecentric proteomics reveals caspase-7
as a substrate of the caspase-1 inflammasomes. Mol. Cell Proteomics 7,
2350-2363 (2008).
490. Fink, S.L. & Cookson, B.T. Caspase-1-dependent pore formation during
pyroptosis leads to osmotic lysis of infected host macrophages. Cell.
Microbiol 8, 1812-1825 (2006).
191

491. Fernandes-Alnemri, T. et al. The pyroptosome: a supramolecular assembly
of ASC dimers mediating inflammatory cell death via caspase-1 activation.
Cell Death Differ 14, 1590-1604 (2007).
492. Kayagaki, N. et al. Non-canonical inflammasome activation targets
caspase-11. Nature 479, 117-121 (2011).
493. Stennicke, H.R. & Salvesen, G.S. Biochemical characteristics of caspases-3,
-6, -7, and -8. J. Biol. Chem 272, 25719-25723 (1997).
494. Fischer, U., Jänicke, R.U. & Schulze-Osthoff, K. Many cuts to ruin: a
comprehensive update of caspase substrates. Cell Death Differ 10, 76-100
(2003).
495. Timmer, J.C. & Salvesen, G.S. Caspase substrates. Cell Death Differ 14, 66-
72 (2007).
496. Gery, I., Gershon, R.K. & Waksman, B.H. Potentiation of the T-lymphocyte
response to mitogens. I. The responding cell. J. Exp. Med 136, 128-142
(1972).
497. Gery, I. & Waksman, B.H. Potentiation of the T-lymphocyte response to
mitogens. II. The cellular source of potentiating mediator(s). J. Exp. Med
136, 143-155 (1972).
498. Gery, I. & Handschumacher, R.E. Potentiation of the T lymphocyte
response to mitogens. III. Properties of the mediator(s) from adherent
cells. Cell. Immunol 11, 162-169 (1974).
499. Rosenwasser, L.J., Dinarello, C.A. & Rosenthal, A.S. Adherent cell function
in murine T-lymphocyte antigen recognition. IV. Enhancement of murine
T-cell antigen recognition by human leukocytic pyrogen. J. Exp. Med 150,
709-714 (1979).
500. March, C.J. et al. Cloning, sequence and expression of two distinct human
interleukin-1 complementary DNAs. Nature 315, 641-647 (1985).
501. Dinarello, C.A. A clinical perspective of IL-1β as the gatekeeper of
inflammation. Eur. J. Immunol 41, 1203-1217 (2011).
502. Dinarello, C.A. IL-18: A TH1-inducing, proinflammatory cytokine and new
member of the IL-1 family. J. Allergy Clin. Immunol 103, 11-24 (1999).
503. Akhter, A. et al. Caspase-7 activation by the Nlrc4/Ipaf inflammasome
restricts Legionella pneumophila infection. PLoS Pathog 5, e1000361
(2009).
192

504. Fernandes-Alnemri, T. & Alnemri, E.S. Assembly, purification, and assay of
the activity of the ASC pyroptosome. Meth. Enzymol 442, 251-270 (2008).
505. Miao, E.A. et al. Caspase-1-induced pyroptosis is an innate immune
effector mechanism against intracellular bacteria. Nat Immunol 11, 1136-
1142 (2010).
506. Delneste, Y. et al. Interferon-gamma switches monocyte differentiation
from dendritic cells to macrophages. Blood 101, 143-150 (2003).
507. Sallusto, F. & Lanzavecchia, A. Efficient presentation of soluble antigen by
cultured human dendritic cells is maintained by granulocyte/macrophage
colony-stimulating factor plus interleukin 4 and downregulated by tumor
necrosis factor alpha. J. Exp. Med 179, 1109-1118 (1994).
508. Zhou, L.J. & Tedder, T.F. CD14+ blood monocytes can differentiate into
functionally mature CD83+ dendritic cells. Proc. Natl. Acad. Sci. U.S.A 93,
2588-2592 (1996).
509. Chapuis, F. et al. Differentiation of human dendritic cells from monocytes
in vitro. Eur. J. Immunol 27, 431-441 (1997).
510. Mohamadzadeh, M. et al. Interleukin 15 skews monocyte differentiation
into dendritic cells with features of Langerhans cells. J. Exp. Med 194,
1013-1020 (2001).
511. Chomarat, P., Banchereau, J., Davoust, J. & Palucka, A.K. IL-6 switches the
differentiation of monocytes from dendritic cells to macrophages. Nat.
Immunol 1, 510-514 (2000).
512. Menetrier-Caux, C. et al. Inhibition of the differentiation of dendritic cells
from CD34(+) progenitors by tumor cells: role of interleukin-6 and
macrophage colony-stimulating factor. Blood 92, 4778-4791 (1998).
513. Allavena, P. et al. IL-10 prevents the generation of dendritic cells from
CD14+ blood monocytes, promotes the differentiation to mature
macrophages and stimulates endocytosis of FITC-dextran. Adv. Exp. Med.
Biol 417, 323-327 (1997).
514. Leff, J.A. et al. Interleukin-1-induced lung neutrophil accumulation and
oxygen metabolite-mediated lung leak in rats. Am. J. Physiol 266, L2-8
(1994).
515. Pugin, J., Ricou, B., Steinberg, K.P., Suter, P.M. & Martin, T.R.
Proinflammatory activity in bronchoalveolar lavage fluids from patients
193

with ARDS, a prominent role for interleukin-1. Am. J. Respir. Crit. Care
Med 153, 1850-1856 (1996).
516. Denis, M. Interferon-gamma-treated murine macrophages inhibit growth
of tubercle bacilli via the generation of reactive nitrogen intermediates.
Cell. Immunol 132, 150-157 (1991).
517. Chan, J., Xing, Y., Magliozzo, R.S. & Bloom, B.R. Killing of virulent
Mycobacterium tuberculosis by reactive nitrogen intermediates produced
by activated murine macrophages. J. Exp. Med 175, 1111-1122 (1992).
518. Breitbach, K. et al. Role of inducible nitric oxide synthase and NADPH
oxidase in early control of Burkholderia pseudomallei infection in mice.
Infect. Immun 74, 6300-6309 (2006).
519. Ouadrhiri, Y., Scorneaux, B., Sibille, Y. & Tulkens, P.M. Mechanism of the
intracellular killing and modulation of antibiotic susceptibility of Listeria
monocytogenes in THP-1 macrophages activated by gamma interferon.
Antimicrob. Agents Chemother 43, 1242-1251 (1999).
520. Miyagi, K., Kawakami, K. & Saito, A. Role of reactive nitrogen and oxygen
intermediates in gamma interferon-stimulated murine macrophage
bactericidal activity against Burkholderia pseudomallei. Infect. Immun 65,
4108-4113 (1997).
521. Arjcharoen, S. et al. Fate of a Burkholderia pseudomallei
lipopolysaccharide mutant in the mouse macrophage cell line RAW 264.7:
possible role for the O-antigenic polysaccharide moiety of
lipopolysaccharide in internalization and intracellular survival. Infect.
Immun 75, 4298-4304 (2007).
522. Breitbach, K., Wongprompitak, P. & Steinmetz, I. Distinct roles for nitric
oxide in resistant C57BL/6 and susceptible BALB/c mice to control
Burkholderia pseudomallei infection. BMC Immunol 12, 20 (2011).
523. Gong, L. et al. The Burkholderia pseudomallei type III secretion system
and BopA are required for evasion of LC3-associated phagocytosis. PLoS
ONE 6, e17852 (2011).
524. Ulett, G.C., Ketheesan, N. & Hirst, R.G. Macrophage-lymphocyte
interactions mediate anti-Burkholderia pseudomallei activity. FEMS
Immunol. Med. Microbiol 21, 283-286 (1998).
194

525. Heinzel, F.P., Sadick, M.D., Holaday, B.J., Coffman, R.L. & Locksley, R.M.
Reciprocal expression of interferon gamma or interleukin 4 during the
resolution or progression of murine leishmaniasis. Evidence for expansion
of distinct helper T cell subsets. J. Exp. Med 169, 59-72 (1989).
526. Heinzel, F.P., Sadick, M.D., Mutha, S.S. & Locksley, R.M. Production of
interferon gamma, interleukin 2, interleukin 4, and interleukin 10 by CD4+
lymphocytes in vivo during healing and progressive murine leishmaniasis.
Proc. Natl. Acad. Sci. U.S.A 88, 7011-7015 (1991).
527. Palma, C., Vendetti, S. & Cassone, A. Role of 4-1BB receptor in the control
played by CD8(+) T cells on IFN-gamma production by Mycobacterium
tuberculosis antigen-specific CD4(+) T Cells. PLoS ONE 5, e11019 (2010).
528. Sester, U. et al. Whole-blood flow-cytometric analysis of antigen-specific
CD4 T-cell cytokine profiles distinguishes active tuberculosis from non-
active states. PLoS ONE 6, e17813 (2011).
529. Tominaga, K. et al. IL-12 synergizes with IL-18 or IL-1beta for IFN-gamma
production from human T cells. Int. Immunol 12, 151-160 (2000).
530. Altés, J. et al. Visceral leishmaniasis: another HIV-associated opportunistic
infection? Report of eight cases and review of the literature. AIDS 5, 201-
207 (1991).
531. Rosenthal, E. et al. HIV and Leishmania coinfection: a review of 91 cases
with focus on atypical locations of Leishmania. Clin. Infect. Dis 31, 1093-
1095 (2000).
532. Lawn, S.D. & Zumla, A.I. Tuberculosis. Lancet 378, 57-72 (2011).
533. Kwan, C.K. & Ernst, J.D. HIV and tuberculosis: a deadly human syndemic.
Clin. Microbiol. Rev 24, 351-376 (2011).
534. Bofill, M. et al. Laboratory control values for CD4 and CD8 T lymphocytes.
Implications for HIV-1 diagnosis. Clin. Exp. Immunol 88, 243-252 (1992).
535. Limmathurotsakul, D. et al. A simple scoring system to differentiate
between relapse and re-infection in patients with recurrent melioidosis.
PLoS Negl Trop Dis 2, e327 (2008).
536. Laws, T.R., Smither, S.J., Lukaszewski, R.A. & Atkins, H.S. Neutrophils are
the predominant cell-type to associate with Burkholderia pseudomallei in a
BALB/c mouse model of respiratory melioidosis. Microb Pathog
(2011).doi:10.1016/j.micpath.2011.07.002
195

537. Egan, A.M. & Gordon, D.L. Burkholderia pseudomallei activates
complement and is ingested but not killed by polymorphonuclear
leukocytes. Infect. Immun 64, 4952-4959 (1996).
538. Chanchamroen, S., Kewcharoenwong, C., Susaengrat, W., Ato, M. &
Lertmemongkolchai, G. Human polymorphonuclear neutrophil responses
to Burkholderia pseudomallei in healthy and diabetic subjects. Infect.
Immun 77, 456-463 (2009).
539. Banerjee, A., Apte, U.M., Smith, R. & Ramaiah, S.K. Higher neutrophil
infiltration mediated by osteopontin is a likely contributing factor to the
increased susceptibility of females to alcoholic liver disease. J. Pathol 208,
473-485 (2006).
540. Wiersinga, W.J. et al. Urokinase receptor is necessary for bacterial defense
against pneumonia-derived septic melioidosis by facilitating phagocytosis.
The Journal of Immunology 184, 3079 -3086 (2010).
541. Brinkmann, V. et al. Neutrophil extracellular traps kill bacteria. Science
303, 1532-1535 (2004).
542. Chan, J.M., Rimm, E.B., Colditz, G.A., Stampfer, M.J. & Willett, W.C.
Obesity, fat distribution, and weight gain as risk factors for clinical diabetes
in men. Diabetes Care 17, 961-969 (1994).
543. Sabanayagam, C. et al. Prevalence of diabetes mellitus, glycemic control,
and associated factors in a Malay population in Singapore. Asia Pac J
Public Health 21, 385-398 (2009).
544. Spiegelman, M. & Marks, H.H. Age and sex variations in the prevalence
and onset of diabetes mellitus. Am J Public Health Nations Health 36, 26-
33 (1946).
545. Danaei, G. et al. National, regional, and global trends in fasting plasma
glucose and diabetes prevalence since 1980: systematic analysis of health
examination surveys and epidemiological studies with 370 country-years
and 2·7 million participants. The Lancet (2011).doi:10.1016/S0140-
6736(11)60679-X
546. Imamura, M. & Maeda, S. Genetics of type 2 diabetes: the GWAS era and
future perspectives. Endocr J (2011).at
196

547. Carpenter, M.W. Gestational diabetes, pregnancy hypertension, and late
vascular disease. Diabetes Care 30 Suppl 2, S246-250 (2007).
548. Papang, R., John, A.S., Abraham, S. & Rao, P.S.S.S. A study of steroid-
induced diabetes mellitus in leprosy. Indian J Lepr 81, 125-129 (2009).
549. Olivarius, N. de F. et al. Patients newly diagnosed with clinical type 2
diabetes during oral glucocorticoid treatment and observed for 14 years:
all-cause mortality and clinical developments. Basic Clin. Pharmacol.
Toxicol 108, 285-288 (2011).
550. Kim, S.Y. et al. Incidence and risk factors of steroid-induced diabetes in
patients with respiratory disease. J. Korean Med. Sci 26, 264-267 (2011).
551. Sharma, A., Pinto Pereira, L.M., Capildeo, K., Charles, K. & Teelucksingh,
S. Steroid-induced iatrogenic disease after treating for
pseudothrombocytopenia. Clin. Appl. Thromb. Hemost 17, 100-102 (2011).
552. Mancini, T., Porcelli, T. & Giustina, A. Treatment of Cushing disease:
overview and recent findings. Ther Clin Risk Manag 6, 505-516 (2010).
553. Cui, Y. & Andersen, D.K. Pancreatogenic diabetes: special considerations
for management. Pancreatology 11, 279-294 (2011).
554. Wilson, C. Diabetes: cystic fibrosis-related diabetes. Nat Rev Endocrinol 7,
375 (2011).
555. Utzschneider, K.M. & Kowdley, K.V. Hereditary hemochromatosis and
diabetes mellitus: implications for clinical practice. Nat Rev Endocrinol 6,
26-33 (2010).
556. Jaeger, M., Aul, C., Söhngen, D., Germing, U. & Schneider, W. Sekundare
Hamochromatose bei polytransfundierten Patienten mit
myelodysplastischen Syndromen. Beitr Infusionsther 30, 464-468 (1992).
557. Cambuli, V.M. et al. Glycometabolic control in acromegalic patients with
diabetes: a study of the effects of different treatments for GH excess and for
hyperglycaemia. J Endocrinol Invest (2011).doi:10.3275/7685
558. Anagnostis, P. et al. Acromegaly: presentation, morbidity and treatment
outcomes at a single centre. Int. J. Clin. Pract 65, 896-902 (2011).
559. McKinlay, J. & Marceau, L. US public health and the 21st century: diabetes
mellitus. The Lancet 356, 757-761 (2000).
197

560. Davis, T.M.E., Pramukkul, P., Suputtamongkol, Y. & Chaowagul, W.
Glucose tolerance in rural diabetic Thais, first-degree relatives and non-
diabetic controls. Diabet Res Clin Prac 27, 171–180 (1995).
561. Muller, L.M.A.J. et al. Increased risk of common infections in patients with
type 1 and type 2 diabetes mellitus. Clin. Infect. Dis 41, 281-288 (2005).
562. Shah, B.R. & Hux, J.E. Quantifying the risk of infectious diseases for people
with diabetes. Diabetes Care 26, 510-513 (2003).
563. Stegenga, M.E. et al. Diabetes does not alter mortality or hemostatic and
inflammatory responses in patients with severe sepsis. Crit. Care Med. 38,
539–45 (2010).
564. Restrepo, B.I. Convergence of the tuberculosis and diabetes epidemics:
renewal of old acquaintances. Clin Infect Dis 45, 436-438 (2007).
565. Joslin, E.P. Treatment of diabetes mellitus: with observations upon the
disease based upon one thousand cases. (Lea & Febiger: Philadelphia,
Pennsylvania, 1916).
566. Bertoni, A.G., Saydah, S. & Brancati, F.L. Diabetes and the risk of infection-
related mortality in the U.S. Diabetes Care 24, 1044-1049 (2001).
567. Chandler, J.R. Malignant external otitis. Laryngoscope 78, 1257-1294
(1968).
568. Ali, T., Meade, K., Anari, S., ElBadawey, M.R. & Zammit-Maempel, I.
Malignant otitis externa: case series. J Laryngol Otol 124, 846-851 (2010).
569. Carfrae, M.J. & Kesser, B.W. Malignant otitis externa. Otolaryngol. Clin.
North Am 41, 537-549, viii-ix (2008).
570. Khaira, A. et al. Retrospective analysis of clinical profile prognostic factors
and outcomes of 19 patients of emphysematous pyelonephritis. Int Urol
Nephrol 41, 959-966 (2009).
571. Michaeli, J., Mogle, P., Perlberg, S., Heiman, S. & Caine, M.
Emphysematous pyelonephritis. J. Urol 131, 203-208 (1984).
572. Ahlering, T.E. et al. Emphysematous pyelonephritis: a 5-year experience
with 13 patients. J. Urol 134, 1086-1088 (1985).
573. Cook, D.J., Achong, M.R. & Dobranowski, J. Emphysematous
pyelonephritis. Complicated urinary tract infection in diabetes. Diabetes
Care 12, 229-232 (1989).
198

574. Esper, A., Moss, M. & Martin, G. The effect of diabetes mellitus on organ
dysfunction with sepsis: an epidemiological study. Crit. Care 13, R18
(2009).
575. Thomsen, R.W. et al. Diabetes and outcome of community-acquired
pneumococcal bacteremia: a 10-year population-based cohort study.
Diabetes Care 27, 70-76 (2004).
576. Tsai, F.-C., Huang, Y.-T., Chang, L.-Y. & Wang, J.-T. Pyogenic liver abscess
as endemic disease, Taiwan. Emerging Infect. Dis 14, 1592-1600 (2008).
577. Gregory, J.E., Golden, A. & Haymaker, W. Mucormycosis of the central
nervous system: a report of three cases. Bull. Johns Hopkins Hosp. 73,
405–419 (1943).
578. Mallis, A., Mastronikolis, S.N., Naxakis, S.S. & Papadas, A.T. Rhinocerebral
mucormycosis: an update. Eur Rev Med Pharmacol Sci 14, 987-992
(2010).
579. Dooley, K.E. & Chaisson, R.E. Tuberculosis and diabetes mellitus:
convergence of two epidemics. Lancet Infect Dis 9, 737-746 (2009).
580. Fine, M.J. et al. Prognosis and outcomes of patients with community-
acquired pneumonia. A meta-analysis. JAMA 275, 134-141 (1996).
581. Thomsen, R.W. et al. Diabetes mellitus as a risk and prognostic factor for
community-acquired bacteremia due to enterobacteria: a 10-year,
population-based study among adults. Clin. Infect. Dis 40, 628-631
(2005).
582. Kornum, J.B. et al. Type 2 diabetes and pneumonia outcomes: a
population-based cohort study. Diabetes Care 30, 2251-2257 (2007).
583. Falagas, M.E., Alexiou, V.G., Giannopoulou, K.P. & Siempos, I.I. Risk
factors for mortality in patients with emphysematous pyelonephritis: a
meta-analysis. J. Urol 178, 880-885; quiz 1129 (2007).
584. Vincent, J.-L., Preiser, J.-C., Sprung, C.L., Moreno, R. & Sakr, Y. Insulin-
treated diabetes is not associated with increased mortality in critically ill
patients. Crit Care 14, R12 (2010).
585. Michalia, M. et al. Diabetes mellitus is an independent risk factor for ICU-
acquired bloodstream infections. Intensive Care Med 35, 448-454 (2009).
586. Tsai, C.L. et al. Impact of diabetes on mortality among patients with
community-acquired bacteremia. J Infect 55, 27–33 (2007).
199

587. McAlister, F.A. et al. The relation between hyperglycemia and outcomes in
2,471 patients admitted to the hospital with community-acquired
pneumonia. Diabetes Care 28, 810-815 (2005).
588. Kaplan, V. et al. Hospitalized community-acquired pneumonia in the
elderly: age- and sex-related patterns of care and outcome in the United
States. Am. J. Respir. Crit. Care Med 165, 766-772 (2002).
589. Cooper, G. & Platt, R. Staphylococcus aureus bacteremia in diabetic
patients. Endocarditis and mortality. Am. J. Med 73, 658-662 (1982).
590. Libert, M. et al. Risk factors for meticillin resistance and outcome of
Staphylococcus aureus bloodstream infection in a Belgian university
hospital. J. Hosp. Infect 68, 17-24 (2008).
591. Carton, J.A. et al. Diabetes mellitus and bacteraemia: a comparative study
between diabetic and non-diabetic patients. Eur J Med 1, 281–87 (1992).
592. Vardakas, K.Z., Siempos, I.I. & Falagas, M.E. Diabetes mellitus as a risk
factor for nosocomial pneumonia and associated mortality. Diabet. Med
24, 1168-1171 (2007).
593. Graham, B.B. et al. Diabetes mellitus does not adversely affect outcomes
from a critical illness. Crit. Care Med 38, 16-24 (2010).
594. Gong, M.N. et al. Clinical predictors of mortality in acute respiratory
distress syndrome: potential role of red cell transfusion. Crit Care Med 33,
1191–1198 (2005).
595. Moss, M. et al. Diabetic patients have a decreased incidence of acute
respiratory distress syndrome. Crit. Care Med 28, 2187-2192 (2000).
596. Watson, R.S. & Angus, D.C. Assessing outcomes in critical care. J Intensive
Care Med 17, 103–111 (2002).
597. Schisterman, E.F., Cole, S.R. & Platt, R.W. Overadjustment bias and
unnecessary adjustment in epidemiologic studies. Epidemiology 20, 488-
495 (2009).
598. Hennekens, C.H. & Buring, J.E. Epidemiology in medicine. (Lippincott,
Williams & Wilkins: Philadelphia, Pennsylvania, 1987).
599. Lassar, O. Ernährungstherapie bei Hautkrankheiten. Dermatologische
Zeitschrift 11, 189 (1904).
200

600. Handmann, E. Über die Ursache der verminderten Resistenz des
Diabetikers gegen Infektionen. Deutsches Archiv für klinische Medizin
102, 1–14 (1911).
601. Da Costa, J.C.J. The opsonic index in diabetes mellitus. A preliminary
record of the findings in 22 cases of glycosuria, with remarks on the
technique of the opsonin test and on its clinical utility. Am J Med Sci 134,
57–70 (1907).
602. Balasoiu, D., van Kessel, K.C., van Kats-Renaud, H.J., Collet, T.J. &
Hoepelman, A.I. Granulocyte function in women with diabetes and
asymptomatic bacteriuria. Diabetes Care 20, 392-395 (1997).
603. Pickup, J.C., Mattock, M.B., Chusney, G.D. & Burt, D. NIDDM as a disease
of the innate immune system: association of acute-phase reactants and
interleukin-6 with metabolic syndrome X. Diabetologia 40, 1286-1292
(1997).
604. Myśliwska, J., Zorena, K., Bakowska, A., Skuratowicz-Kubica, A. &
Myśliwski, A. Significance of tumor necrosis factor alpha in patients with
long-standing type-I diabetes mellitus. Horm. Metab. Res 30, 158-161
(1998).
605. Zozuliñska, D., Majchrzak, A., Sobieska, M., Wiktorowicz, K. & Wierusz-
Wysocka, B. Serum interleukin-8 level is increased in diabetic patients.
Diabetologia 42, 117-118 (1999).
606. Yende, S. et al. The influence of pre-existing diabetes mellitus on the host
immune response and outcome of pneumonia: analysis of two multicentre
cohort studies. Thorax 65, 870-877 (2010).
607. Rao, A.K., Chouhan, V., Chen, X., Sun, L. & Boden, G. Activation of the
tissue factor pathway of blood coagulation during prolonged hyperglycemia
in young healthy men. Diabetes 48, 1156-1161 (1999).
608. Stegenga, M.E. et al. Hyperglycemia stimulates coagulation, whereas
hyperinsulinemia impairs fibrinolysis in healthy humans. Diabetes 55,
1807-1812 (2006).
609. Geerlings, S.E. et al. Cytokine secretion is impaired in women with diabetes
mellitus. Eur. J. Clin. Invest 30, 995-1001 (2000).
610. Aird, W.C. The role of the endothelium in severe sepsis and multiple organ
dysfunction syndrome. Blood 101, 3765-3777 (2003).
201

611. Schuetz, P. et al. Influence of diabetes on endothelial cell response during
sepsis. Diabetologia 54, 996-1003 (2011).
612. Neelamegham, S., Taylor, A.D., Burns, A.R., Smith, C.W. & Simon, S.I.
Hydrodynamic shear shows distinct roles for LFA-1 and Mac-1 in
neutrophil adhesion to intercellular adhesion molecule-1. Blood 92, 1626-
1638 (1998).
613. Hentzen, E.R. et al. Sequential binding of CD11a/CD18 and CD11b/CD18
defines neutrophil capture and stable adhesion to intercellular adhesion
molecule-1. Blood 95, 911-920 (2000).
614. Li, J., Jin, H.-B., Sun, Y.-M., Su, Y. & Wang, L.-F. KB-R7943 inhibits high
glucose-induced endothelial ICAM-1 expression and monocyte-endothelial
adhesion. Biochem. Biophys. Res. Commun 392, 516-519 (2010).
615. Takami, S., Yamashita, S., Kihara, S., Kameda-Takemura, K. & Matsuzawa,
Y. High concentration of glucose induces the expression of intercellular
adhesion molecule-1 in human umbilical vein endothelial cells.
Atherosclerosis 138, 35-41 (1998).
616. Altannavch, T.S., Roubalová, K., Kucera, P. & Andel, M. Effect of high
glucose concentrations on expression of ELAM-1, VCAM-1 and ICAM-1 in
HUVEC with and without cytokine activation. Physiol Res 53, 77-82
(2004).
617. Andersen, B., Goldsmith, G.H. & Spagnuolo, P.J. Neutrophil adhesive
dysfunction in diabetes mellitus; the role of cellular and plasma factors. J.
Lab. Clin. Med 111, 275-285 (1988).
618. Delamaire, M. et al. Impaired leucocyte functions in diabetic patients.
Diabet. Med 14, 29-34 (1997).
619. Morigi, M. et al. Leukocyte-endothelial interaction is augmented by high
glucose concentrations and hyperglycemia in a NF-κB-dependent fashion.
J. Clin. Invest 101, 1905-1915 (1998).
620. Taki, H., Kashiwagi, A., Tanaka, Y. & Horiike, K. Expression of intercellular
adhesion molecules 1 (ICAM-1) via an osmotic effect in human umbilical
vein endothelial cells exposed to high glucose medium. Life Sci 58, 1713-
1721 (1996).
202

621. Park, C.W. et al. High glucose-induced intercellular adhesion molecule-1
(ICAM-1) expression through an osmotic effect in rat mesangial cells is
PKC-NF-kappa B-dependent. Diabetologia 43, 1544-1553 (2000).
622. Boyden, S. The chemotactic effect of mixtures of antibody and antigen on
polymorphonuclear leucocytes. J. Exp. Med 115, 453-466 (1962).
623. Mowat, A. & Baum, J. Chemotaxis of polymorphonuclear leukocytes from
patients with diabetes mellitus. N. Engl. J. Med 284, 621-627 (1971).
624. Valerius, N.H. et al. Neutrophil and lymphocyte function in patients with
diabetes mellitus. Acta Med Scand 211, 463-467 (1982).
625. Nelson, R.D., Quie, P.G. & Simmons, R.L. Chemotaxis under agarose: a
new and simple method for measuring chemotaxis and spontaneous
migration of human polymorphonuclear leukocytes and monocytes. J.
Immunol 115, 1650-1656 (1975).
626. Tater, D., Tepaut, B., Bercovici, J.P. & Youinou, P. Polymorphonuclear cell
derangements in type I diabetes. Horm. Metab. Res 19, 642-647 (1987).
627. Bybee, J.D. & Rogers, D.E. The phagocytic activity of polymorphonuclear
leukocytes obtained from patients with diabetes mellitus. J. Lab. Clin. Med
64, 1-13 (1964).
628. Bagdade, J.D., Root, R.K. & Bulger, R.J. Impaired leukocyte function in
patients with poorly controlled diabetes. Diabetes 23, 9-15 (1974).
629. Bagdade, J.D. Phagocytic and microbicidal function in diabetes mellitus.
Acta Endocrinol Suppl (Copenh) 205, 27-34 (1976).
630. Alexiewicz, J.M., Kumar, D., Smogorzewski, M., Klin, M. & Massry, S.G.
Polymorphonuclear leukocytes in non-insulin-dependent diabetes mellitus:
abnormalities in metabolism and function. Ann. Intern. Med 123, 919-924
(1995).
631. Davidson, N.J., Sowden, J.M. & Fletcher, J. Defective phagocytosis in
insulin controlled diabetics: evidence for a reaction between glucose and
opsonising proteins. J. Clin. Pathol 37, 783-786 (1984).
632. Dziatkowiak, H., Kowalska, M. & Denys, A. Phagocytic and bactericidal
activity of granulocytes in diabetic children. Diabetes 31, 1041-1043 (1982).
633. Tan, J.S., Anderson, J.L., Watanakunakorn, C. & Phair, J.P. Neutrophil
dysfunction in diabetes mellitus. J. Lab. Clin. Med 85, 26-33 (1975).
203

634. Bagdade, J.D., Nielson, K.L. & Bulger, R.J. Reversible abnormalities in
phagocytic function in poorly controlled diabetic patients. Am. J. Med. Sci
263, 451-456 (1972).
635. Wilson, R.M. & Reeves, W.G. Neutrophil phagocytosis and killing in
insulin-dependent diabetes. Clin. Exp. Immunol 63, 478-484 (1986).
636. Shah, S.V., Wallin, J.D. & Eilen, S.D. Chemiluminescence and superoxide
anion production by leukocytes from diabetic patients. J. Clin. Endocrinol.
Metab 57, 402-409 (1983).
637. Wykretowicz, A., Wierusz-Wysocka, B., Wysocki, J., Szczepanik, A. &
Wysocki, H. Impairment of the oxygen-dependent microbicidal
mechanisms of polymorphonuclear neutrophils in patients with type 2
diabetes is not associated with increased susceptibility to infection.
Diabetes Res. Clin. Pract 19, 195-201 (1993).
638. Sato, N., Kashima, K., Tanaka, Y., Shimizu, H. & Mori, M. Effect of
granulocyte-colony stimulating factor on generation of oxygen-derived free
radicals and myeloperoxidase activity in neutrophils from poorly controlled
NIDDM patients. Diabetes 46, 133-137 (1997).
639. Daoud, A.K., Tayyar, M.A., Fouda, I.M. & Harfeil, N.A. Effects of diabetes
mellitus vs. in vitro hyperglycemia on select immune cell functions. J
Immunotoxicol 6, 36-41 (2009).
640. Grinberg, N., Elazar, S., Rosenshine, I. & Shpigel, N.Y. β-Hydroxybutyrate
abrogates formation of bovine neutrophil extracellular traps and
bactericidal activity against mammary pathogenic Escherichia coli. Infect
Immun 76, 2802-2807 (2008).
641. Hill, H.R., Augustine, N.H., Rallison, M.L. & Santos, J.I. Defective
monocyte chemotactic responses in diabetes mellitus. J. Clin. Immunol 3,
70-77 (1983).
642. Geisler, C., Almdal, T., Bennedsen, J., Rhodes, J.M. & Kølendorf, K.
Monocyte functions in diabetes mellitus. Acta Pathol Microbiol Immunol
Scand C 90, 33-37 (1982).
643. Katz, S., Klein, B., Elian, I., Fishman, P. & Djaldetti, M. Phagocytotic
activity of monocytes from diabetic patients. Diabetes Care 6, 479-482
(1983).
204

644. Nandy, D., Janardhanan, R., Mukhopadhyay, D. & Basu, A. Effect of
hyperglycemia on human monocyte activation. J Investig Med
(2011).doi:10.231/JIM.0b013e31820ee432
645. Rattan, V., Shen, Y., Sultana, C., Kumar, D. & Kalra, V.K. Glucose-induced
transmigration of monocytes is linked to phosphorylation of PECAM-1 in
cultured endothelial cells. Am. J. Physiol 271, E711-717 (1996).
646. Kitahara, M., Eyre, H.J., Lynch, R.E., Rallison, M.L. & Hill, H.R. Metabolic
activity of diabetic monocytes. Diabetes 29, 251-256 (1980).
647. Foss-Freitas, M.C., Foss, N.T., Donadi, E.A. & Foss, M.C. In vitro TNF-
alpha and IL-6 production by adherent peripheral blood mononuclear cells
obtained from type 1 and type 2 diabetic patients evaluated according to
the metabolic control. Ann. N. Y. Acad. Sci 1079, 177-180 (2006).
648. MacCuish, A.C., Urbaniak, S.J., Campbell, C.J., Duncan, L.J. & Irvine, W.J.
Phytohemagglutinin transformation and circulating lymphocyte
subpopulations in insulin-dependent diabetic patients. Diabetes 23, 708-
712 (1974).
649. Speert, D.P. & Silva, J. Abnormalities of in vitro lymphocyte response to
mitogens in diabetic children during acute ketoacidosis. Am. J. Dis. Child
132, 1014-1017 (1978).
650. Eibl, N. et al. Impaired primary immune response in type-1 diabetes:
results from a controlled vaccination study. Clin. Immunol 103, 249-259
(2002).
651. Spatz, M. et al. Impaired primary immune response in type-1 diabetes.
Functional impairment at the level of APCs and T-cells. Cell. Immunol 221,
15-26 (2003).
652. Casey, J. & Sturm, C. Impaired response of lymphocytes from non-insulin-
dependent diabetics to staphage lysate and tetanus antigen. J. Clin.
Microbiol 15, 109-114 (1982).
653. Casey, J.I., Heeter, B.J. & Klyshevich, K.A. Impaired response of
lymphocytes of diabetic subjects to antigen of Staphylococcus aureus. J.
Infect. Dis 136, 495-501 (1977).
654. Da Costa, J.C.J. & Beardsley, E.J.G. The resistance of diabetics to bacterial
infection: a study of the opsonophagocytic properties of the blood in 74
205

cases of diabetes mellitus and related conditions. Am. J. Med. Sci. 136,
361–373 (1908).
655. Liberatore, R.D.R. et al. Is immunity in diabetic patients influencing the
susceptibility to infections? Immunoglobulins, complement and phagocytic
function in children and adolescents with type 1 diabetes mellitus. Pediatr
Diabetes 6, 206-212 (2005).
656. Moen, J.K. & Reimann, H.A. Immune reactions in diabetes. Arch Intern
Med 51, 789-795 (1933).
657. Richardson, R. Immunity in diabetes: influence of diabetes on the
development of antibacterial properties in blood. J. Clin. Invest. 12, 1143-
1149 (1933).
658. Fiçicioğlu, C. et al. Reduced immune response to hepatitis B vaccine in
children with insulin dependent diabetes. Acta Paediatr Jpn 37, 687-690
(1995).
659. Pozzilli, P. et al. Reduced protection against hepatitis B virus following
vaccination in patients with type 1 (insulin-dependent) diabetes.
Diabetologia 30, 817-819 (1987).
660. Alavian, S.-M. & Tabatabaei, S.V. The effect of diabetes mellitus on
immunological response to hepatitis B virus vaccine in individuals with
chronic kidney disease: A meta-analysis of current literature. Vaccine 28,
3773-3777 (2010).
661. Bouter, K.P. et al. Humoral immune response to a yeast-derived hepatitis B
vaccine in patients with type 1 diabetes mellitus. Diabet. Med 9, 66-69
(1992).
662. Douvin, C. et al. Hepatitis B vaccination in diabetic patients. Randomized
trial comparing recombinant vaccines containing and not containing pre-
S2 antigen. Diabetes Care 20, 148-151 (1997).
663. Brydak, L.B. & Machala, M. Humoral immune response to influenza
vaccination in patients from high risk groups. Drugs 60, 35-53 (2000).
664. Pozzilli, P. et al. The immune response to influenza vaccination in diabetic
patients. Diabetologia 29, 850-854 (1986).
665. Muszkat, M. et al. Response to influenza vaccination in community and in
nursing home residing elderly: relation to clinical factors. Exp. Gerontol
38, 1199-1203 (2003).
206

666. Tamer, A. et al. Impaired immunity against tetanus in type 2 diabetes.
Med. Sci. Monit 11, CR580-584 (2005).
667. Kiliç, D. et al. Seroprevalence of tetanus immunity among noninsulin-
dependent diabetes mellitus patients. J. Diabetes Complicat 17, 258-263
(2003).
668. Beam, T.R., Crigler, E.D., Goldman, J.K. & Schiffman, G. Antibody
response to polyvalent pneumococcal polysaccharide vaccine in diabetics.
JAMA 244, 2621-2624 (1980).
669. Mijovic, C.H., Fletcher, J.A., Bradwell, A.R. & Barnett, A.H. Low C4 levels
in type 1 (insulin-dependent) diabetes. Diabetologia 30, 824 (1987).
670. Vergani, D., Johnston, C., B-Abdullah, N. & Barnett, A.H. Low serum C4
concentrations: an inherited predisposition to insulin dependent diabetes?
Br Med J (Clin Res Ed) 286, 926-928 (1983).
671. Jenhani, F., Bardi, R., Gorgi, Y., Ayed, K. & Jeddi, M. C4 polymorphism in
multiplex families with insulin dependent diabetes in the Tunisian
population: standard C4 typing methods and RFLP analysis. J. Autoimmun
5, 149-160 (1992).
672. Hernández-Mijares, A. et al. Levels of C3 in patients with severe, morbid
and extreme obesity: its relationship to insulin resistance and different
cardiovascular risk factors. Int J Obes (Lond) 31, 927-932 (2007).
673. Karlsson, E., Shaat, N. & Groop, L. Can complement factors 5 and 8 and
transthyretin be used as biomarkers for MODY 1 (HNF4A-MODY) and
MODY 3 (HNF1A-MODY)? Diabet. Med 25, 788-791 (2008).
674. Davin, J.C. et al. IgG glycation and function during continuous ambulatory
peritoneal dialysis. Nephrol. Dial. Transplant 12, 310-314 (1997).
675. Hostetter, M.K. Handicaps to host defense. Effects of hyperglycemia on C3
and Candida albicans. Diabetes 39, 271-275 (1990).
676. Williams, N.L., Morris, J.L., Rush, C., Govan, B.L. & Ketheesan, N. Impact
of streptozotocin-induced diabetes on functional responses of dendritic
cells and macrophages toward Burkholderia pseudomallei. FEMS Immunol
Med Microbiol (2010).doi:10.1111/j.1574-695X.2010.00767.x
677. Hodgson, K.A., Morris, J.L., Feterl, M.L., Govan, B.L. & Ketheesan, N.
Altered macrophage function is associated with severe Burkholderia
207

pseudomallei infection in a murine model of type 2 diabetes. Microbes
Infect (2011).doi:10.1016/j.micinf.2011.07.008
678. Coleman, D.L. & Hummel, K.P. Studies with the mutation, diabetes, in the
mouse. Diabetologia 3, 238-248 (1967).
679. Chen, H. et al. Evidence that the diabetes gene encodes the leptin receptor:
identification of a mutation in the leptin receptor gene in db/db mice. Cell
84, 491-495 (1996).
680. Park, S., Rich, J., Hanses, F. & Lee, J.C. Defects in innate immunity
predispose C57BL/6J-Leprdb/Leprdb mice to infection by Staphylococcus
aureus. Infect. Immun 77, 1008-1014 (2009).
681. Limmathurotsakul, D. et al. Role and significance of quantitative urine
cultures in diagnosis of melioidosis. J. Clin. Microbiol 43, 2274-2276
(2005).
682. Nelder, J.A. & Wedderburn, R.W.M. Generalized linear models. Journal of
the Royal Statistical Society 135, 370 (1972).
683. Verhulst, P.F. Notice sur la loi que la population suit dans son
accroissement. Correspondance Mathématique et Physique, publiée par A.
Quetelet 4, 113–121 (1838).
684. Verhulst, P.F. Recherches mathématiques sur la loi d’accroissement de la
population. Nouveaux Mémoires de l’Académie Royale des Sciences, des
Lettres et des Beaux-Arts de Belgique 18, 1–32 (1845).
685. Berkson, J. Application of the logistic function to bio-assay. Journal of the
American Statistical Association 39, 357–365 (1944).
686. Cramer, J.S. Chapter 9: The origins and development of the logit model.
Logit Models from Economics and Other Fields 149–157 (2003).
687. Cox, D.R. Analysis of Binary Data. (Chapman & Hall: London, England,
1969).
688. Hosmer, D.W. & Lemeshow, S. Applied Logistic Regression. (Wiley: New
York, USA, 1989).
689. Peduzzi, P., Concato, J., Kemper, E., Holford, T.R. & Feinstein, A.R. A
simulation study of the number of events per variable in logistic regression
analysis. J Clin Epidemiol 49, 1373-1379 (1996).
208

690. Vittinghoff, E. & McCulloch, C.E. Relaxing the rule of ten events per
variable in logistic and Cox regression. Am. J. Epidemiol 165, 710-718
(2007).
691. Good, P.I. & Hardin, J.W. Common Errors in Statistics. (John Wiley &
Sons: Hoboken, New Jersey, USA, 2009).
692. Rothman, K.J. Modern Epidemiology. (Lippincott Williams and Wilkins:
1986).
693. Robins, J.M. Data, design, and background knowledge in etiologic
inference. Epidemiology 12, 313-320 (2001).
694. Pearl, J. Chapter 6.2. Why there is no statistical test for confounding, Why
many think there is, and Why they are almost right. Causality: Models,
Reasoning and Inference 182f (2009).
695. Katz, M.H. Multivariable analysis: a primer for readers of medical research.
Ann. Intern. Med 138, 644-650 (2003).
696. Cole, S.R. et al. Illustrating bias due to conditioning on a collider. Int J
Epidemiol 39, 417-420 (2010).
697. Maldonado, G. & Greenland, S. Simulation study of confounder-selection
strategies. Am. J. Epidemiol 138, 923-936 (1993).
698. Greenland, S., Pearl, J. & Robins, J.M. Causal diagrams for epidemiologic
research. Epidemiology 10, 37-48 (1999).
699. Hernán, M.A., Hernández-Díaz, S., Werler, M.M. & Mitchell, A.A. Causal
knowledge as a prerequisite for confounding evaluation: an application to
birth defects epidemiology. Am. J. Epidemiol 155, 176-184 (2002).
700. Victora, C.G., Huttly, S.R., Fuchs, S.C. & Olinto, M.T.A. The role of
conceptual frameworks in epidemiological analysis: a hierarchical
approach. Int J Epidemiol 26, 224-227 (1997).
701. Kumar, A. et al. Duration of hypotension before initiation of effective
antimicrobial therapy is the critical determinant of survival in human
septic shock. Crit. Care Med 34, 1589-1596 (2006).
702. Buckley, J.F., Singer, M. & Clapp, L.H. Role of KATP channels in sepsis.
Cardiovasc. Res 72, 220-230 (2006).
703. Warrillow, S., Egi, M. & Bellomo, R. Randomized, double-blind, placebo-
controlled crossover pilot study of a potassium channel blocker in patients
with septic shock. Crit. Care Med. 34, 980–985 (2006).
209

704. Morelli, A. et al. Glibenclamide dose response in patients with septic shock:
effects on norepinephrine requirements, cardiopulmonary performance,
and global oxygen transport. Shock 28, 530–535 (2007).
705. Cole, S.R. & Hernán, M.A. Fallibility in estimating direct effects.
International Journal of Epidemiology 31, 163–165 (2002).
706. McNicholas, C.M. et al. Sensitivity of a renal K+ channel (ROMK2) to the
inhibitory sulfonylurea compound glibenclamide is enhanced by
coexpression with the ATP-binding cassette transporter cystic fibrosis
transmembrane regulator. Proc. Natl. Acad. Sci. U.S.A 93, 8083-8088
(1996).
707. Payen, L. et al. The sulphonylurea glibenclamide inhibits multidrug
resistance protein (MRP1) activity in human lung cancer cells. Br. J.
Pharmacol 132, 778-784 (2001).
708. Melin, P., Hosy, E., Vivaudou, M. & Becq, F. CFTR inhibition by
glibenclamide requires a positive charge in cytoplasmic loop three.
Biochim. Biophys. Acta 1768, 2438-2446 (2007).
709. Southern, E.M. Detection of specific sequences among DNA fragments
separated by gel electrophoresis. J. Mol. Biol. 98, 503-517 (1975).
710. Augenlicht, L.H. & Kobrin, D. Cloning and screening of sequences
expressed in a mouse colon tumor. Cancer Res 42, 1088-1093 (1982).
711. Augenlicht, L.H. et al. Expression of cloned sequences in biopsies of human
colonic tissue and in colonic carcinoma cells induced to differentiate in
vitro. Cancer Res 47, 6017-6021 (1987).
712. Illumina HumanWG-6 v3.0 and HumanRef-8 v3.0 Expression BeadChips.
(2008).
713. Chapter 18, The Reference Sequence (RefSeq) Project. The NCBI
handbook (2002).at
714. Chapter 21, UniGene: a unified view of the transcriptome. The NCBI
Handbook (2002).at
715. Fare, T.L. et al. Effects of atmospheric ozone on microarray data quality.
Anal. Chem 75, 4672-4675 (2003).
716. Bonferroni, C.E. Il calcolo delle assicurazioni su gruppi di teste. Studi in
Onore del Professore Salvatore Ortu Carboni 13–60 (1935).
210

717. Dunn, O.J. Multiple comparisons among means. Journal of the American
Statistical Association 56, 52–64 (1961).
718. Holm, S. A simple sequentially rejective multiple test procedure.
Scandinavian Journal of Statistics 6, 65-70 (1979).
719. Benjamini, Y. & Hochberg, Y. Controlling the false discovery rate: a
practical and powerful approach to multiple testing. J R Statist Soc B 57,
289–300 (1995).
720. Perneger, T.V. What’s wrong with Bonferroni adjustments. BMJ 316,
1236–1238 (1998).
721. Sterne, J.A. & Davey Smith, G. Sifting the evidence-what’s wrong with
significance tests? BMJ 322, 226-231 (2001).
722. Keppel, G. & Wickens, T.D. Design and Analysis: A Researcher’s
Handbook. (Prentice Hall: Upper Saddle River, New Jersey, 2004).
723. Shi, L. et al. The MicroArray Quality Control (MAQC) project shows inter-
and intraplatform reproducibility of gene expression measurements. Nat.
Biotechnol 24, 1151-1161 (2006).
724. Subramanian, A. et al. Gene set enrichment analysis: A knowledge-based
approach for interpreting genome-wide expression profiles. Proc. Natl.
Acad. Sci. U.S.A 102, 15545 -15550 (2005).
725. Lynn, D. et al. Curating the innate immunity interactome. BMC Systems
Biology 4, 117 (2010).
726. Croft, D. et al. Reactome: a database of reactions, pathways and biological
processes. Nucleic Acids Res 39, D691-697 (2011).
727. Rodbell, M. The role of hormone receptors and GTP-regulatory proteins in
membrane transduction. Nature 284, 17-22 (1980).
728. The Nobel Prize in Physiology or Medicine 1994. (2011).at
729. Wu, G., Feng, X. & Stein, L. A human functional protein interaction
network and its application to cancer data analysis. Genome Biol 11, R53
(2010).
730. Kanehisa, M., Goto, S., Kawashima, S., Okuno, Y. & Hattori, M. The KEGG
resource for deciphering the genome. Nucleic Acids Res 32, D277-280
(2004).
211

731. Ashburner, M. et al. Gene ontology: tool for the unification of biology. The
Gene Ontology Consortium. Nat. Genet 25, 25-29 (2000).
732. Brazma, A. et al. Minimum information about a microarray experiment
(MIAME)-toward standards for microarray data. Nat. Genet 29, 365-371
(2001).
733. Ioannidis, J.P.A. et al. Repeatability of published microarray gene
expression analyses. Nat. Genet 41, 149-155 (2009).
734. McCane, D.R. et al. Comparison of tests for glycated haemoglobin and
fasting and two hour plasma glucose concentrations as diagnostic methods
for diabetes. BMJ 308, 1323-1328 (1994).
735. Lamkanfi, M. et al. Glyburide inhibits the Cryopyrin/Nalp3 inflammasome.
J. Cell Biol 187, 61-70 (2009).
736. Meduri, G.U. et al. Persistent elevation of inflammatory cytokines predicts
a poor outcome in ARDS. Plasma IL-1 beta and IL-6 levels are consistent
and efficient predictors of outcome over time. Chest 107, 1062-1073 (1995).
737. Kikkawa, T. et al. Assessment of IL-18 values in septic acute lung
injury/acute respiratory distress syndrome patients. Critical Care 13, P368
(2009).
738. Koh, Y., Hybertson, B.M., Jepson, E.K., Cho, O.J. & Repine, J.E. Cytokine-
induced neutrophil chemoattractant is necessary for interleukin-1-induced
lung leak in rats. J. Appl. Physiol 79, 472-478 (1995).
739. Ware, L.B. & Matthay, M.A. The acute respiratory distress syndrome. N.
Engl. J. Med 342, 1334-1349 (2000).
740. Pompermayer, K. et al. The ATP-sensitive potassium channel blocker
glibenclamide prevents renal ischemia/reperfusion injury in rats. Kidney
Int 67, 1785-1796 (2005).
741. Pompermayer, K. et al. Effects of the treatment with glibenclamide, an
ATP-sensitive potassium channel blocker, on intestinal ischemia and
reperfusion injury. Eur. J. Pharmacol 556, 215-222 (2007).
742. Riddle, M.C. Sulfonylureas differ in effects on ischemic preconditioning—Is
it time to retire glyburide? J Clin Endocrinol Metab 88, 528-530 (2003).
743. Muruve, D.A. et al. The inflammasome recognizes cytosolic microbial and
host DNA and triggers an innate immune response. Nature 452, 103-107
(2008).
212

744. Haskó, G., Deitch, E.A., Németh, Z.H., Kuhel, D.G. & Szabó, C. Inhibitors of
ATP-binding cassette transporters suppress interleukin-12 p40 production
and major histocompatibility complex II up-regulation in macrophages. J.
Pharmacol. Exp. Ther 301, 103-110 (2002).
745. Hamon, Y. et al. ABC1 promotes engulfment of apoptotic cells and
transbilayer redistribution of phosphatidylserine. Nat. Cell Biol 2, 399-406
(2000).
746. Di Virgillo, F. Liaisons dangereuses: P2X(7) and the inflammasome.
Trends Pharmacol. Sci. 28, 465–72 (2007).
747. Jiang, H. et al. Stimulation of rat erythrocyte P2X7 receptor induces the
release of epoxyeicosatrienoic acids. Br. J. Pharmacol. 151, 1033–40
(2007).
748. Dal-Secco, D. et al. Hydrogen sulfide augments neutrophil migration
through enhancement of adhesion molecule expression and prevention of
CXCR2 internalization: role of ATP-sensitive potassium channels. J.
Immunol. 181, 4287–98 (2008).
749. Pettitt, S.J. et al. Agouti C57BL/6N embryonic stem cells for mouse genetic
resources. Nat. Methods 6, 493-495 (2009).
750. Skarnes, W.C. et al. A conditional knockout resource for the genome-wide
study of mouse gene function. Nature 474, 337-342 (2011).
751. Leakey, A.K., Ulett, G.C. & Hirst, R.G. BALB/c and C57Bl/6 mice infected
with virulent Burkholderia pseudomallei provide contrasting animal
models for the acute and chronic forms of human melioidosis. Microb.
Pathog 24, 269-275 (1998).
752. Hoppe, I. et al. Characterization of a murine model of melioidosis:
comparison of different strains of mice. Infect. Immun 67, 2891-2900
(1999).
753. Conejero, L. et al. Low-dose exposure of C57BL/6 mice to Burkholderia
pseudomallei mimics chronic human melioidosis. Am. J. Pathol 179, 270-
280 (2011).
754. Reiner, S.L. & Locksley, R.M. The regulation of immunity to Leishmania
major. Annu. Rev. Immunol 13, 151-177 (1995).
213

755. Chackerian, A.A. & Behar, S.M. Susceptibility to Mycobacterium
tuberculosis: lessons from inbred strains of mice. Tuberculosis (Edinb) 83,
279-285 (2003).
756. Autenrieth, I.B., Beer, M., Bohn, E., Kaufmann, S.H. & Heesemann, J.
Immune responses to Yersinia enterocolitica in susceptible BALB/c and
resistant C57BL/6 mice: an essential role for gamma interferon. Infect.
Immun 62, 2590-2599 (1994).
757. Mohammadi, M., Redline, R., Nedrud, J. & Czinn, S. Role of the host in
pathogenesis of Helicobacter-associated gastritis: H. felis infection of
inbred and congenic mouse strains. Infect. Immun. 64, 238-245 (1996).
758. Morissette, C., Skamene, E. & Gervais, F. Endobronchial inflammation
following Pseudomonas aeruginosa infection in resistant and susceptible
strains of mice. Infect. Immun. 63, 1718-1724 (1995).
759. Ulett, G.C., Ketheesan, N. & Hirst, R.G. Cytokine gene expression in
innately susceptible BALB/c mice and relatively resistant C57BL/6 mice
during infection with virulent Burkholderia pseudomallei. Infect. Immun
68, 2034-2042 (2000).
760. Barnes, J.L. et al. Induction of multiple chemokine and colony-stimulating
factor genes in experimental Burkholderia pseudomallei infection.
Immunol. Cell Biol 79, 490-501 (2001).
761. Ralph, P. & Nakoinz, I. Phagocytosis and cytolysis by a macrophage tumour
and its cloned cell line. Nature 257, 393-394 (1975).
762. Watanabe, H., Numata, K., Ito, T., Takagi, K. & Matsukawa, A. Innate
immune response in Th1- and Th2-dominant mouse strains. Shock 22,
460-466 (2004).
763. Mosmann, T.R., Cherwinski, H., Bond, M.W., Giedlin, M.A. & Coffman,
R.L. Two types of murine helper T cell clone. I. Definition according to
profiles of lymphokine activities and secreted proteins. J. Immunol 136,
2348-2357 (1986).
764. Atkins, T. et al. A mutant of Burkholderia pseudomallei, auxotrophic in the
branched chain amino acid biosynthetic pathway, is attenuated and
protective in a murine model of melioidosis. Infect Immun 70, 5290-5294
(2002).
214

765. Wiersinga, W.J. et al. Inflammation patterns induced by different
Burkholderia species in mice. Cell. Microbiol. 10, 81-87 (2008).
766. van Schaik, E., Tom, M., DeVinney, R. & Woods, D.E. Development of
novel animal infection models for the study of acute and chronic
Burkholderia pseudomallei pulmonary infections. Microbes Infect 10,
1291-1299 (2008).
767. Cuccui, J. et al. Development of signature-tagged mutagenesis in
Burkholderia pseudomallei to identify genes important in survival and
pathogenesis. Infect. Immun 75, 1186-1195 (2007).
768. Nelson, M. et al. Development of an acute model of inhalational
melioidosis in the common marmoset (Callithrix jacchus). Int J Exp Pathol
92, 428-435 (2011).
769. Manzeniuk, I.N. et al. [Burkholderia mallei and Burkholderia
pseudomallei. Study of immuno- and pathogenesis of glanders and
melioidosis. Heterologous vaccines]. Antibiot. Khimioter 44, 21-26 (1999).
770. Narita, M., Loganathan, P., Hussein, A., Jamaluddin, A. & Joseph, P.G.
Pathological changes in goats experimentally inoculated with Pseudomonas
pseudomallei. Natl Inst Anim Health Q (Tokyo) 22, 170-179 (1982).
771. Najdenski, H., Kussovski, V. & Vesselinova, A. Experimental Burkholderia
pseudomallei infection of pigs. J. Vet. Med. B Infect. Dis. Vet. Public Health
51, 225-230 (2004).
772. O’Quinn, A.L., Wiegand, E.M. & Jeddeloh, J.A. Burkholderia pseudomallei
kills the nematode Caenorhabditis elegans using an endotoxin-mediated
paralysis. Cell. Microbiol 3, 381-393 (2001).
773. Gan, Y.-H. et al. Characterization of Burkholderia pseudomallei infection
and identification of novel virulence factors using a Caenorhabditis elegans
host system. Mol. Microbiol 44, 1185-1197 (2002).
774. von Mering, J.F. & Minkowski, O. Diabetes mellitus nach
Pankreasexstirpation. Archiv für experimentelle Pathologie und
Pharmakologie 26, 371–387 (1890).
775. Banting, F.G. & Best, C.H. The internal secretion of the pancreas. J. Lab.
Clin. Invest. 7, 251–266 (1922).
776. Makino, S. et al. Breeding of a non-obese, diabetic strain of mice. Jikken
Dobutsu 29, 1-13 (1980).
215

777. Makino, S., Hayashi, Y. & Muraoka, Y. Establishment of the nonobese-
diabetic (NOD) mouse model. Current topics in clinical and experimental
aspects of diabetes mellitus 25–52 (1985).
778. Bowman, M.A., Leiter, E.H. & Atkinson, M.A. Prevention of diabetes in the
NOD mouse: implications for therapeutic intervention in human disease.
Immunol. Today 15, 115-120 (1994).
779. Harada, M. & Makino, S. Promotion of spontaneous diabetes in non-obese
diabetes-prone mice by cyclophosphamide. Diabetologia 27, 604-606
(1984).
780. Hattori, M. et al. A single recessive non-MHC diabetogenic gene
determines the development of insulitis in the presence of an MHC-linked
diabetogenic gene in NOD mice. J. Autoimmun. 3, 1-10 (1990).
781. Mathews, C.E., Graser, R.T., Serreze, D.V. & Leiter, E.H. Reevaluation of
the major histocompatibility complex genes of the NOD-progenitor
CTS/Shi strain. Diabetes 49, 131-134 (2000).
782. Yoshioka, M., Kayo, T., Ikeda, T. & Koizumi, A. A novel locus, Mody4, distal
to D7Mit189 on chromosome 7 determines early-onset NIDDM in
nonobese C57BL/6 (Akita) mutant mice. Diabetes 46, 887-894 (1997).
783. Soares, M.B. et al. RNA-mediated gene duplication: the rat preproinsulin I
gene is a functional retroposon. Mol. Cell. Biol. 5, 2090-2103 (1985).
784. Mathews, C.E. & Leiter, E.H. Chapter 18: Rodent models for the study of
diabetes. Joslin’s Diabetes Mellitus (2005).
785. Coleman, D.L. & Eicher, E.M. Fat (fat) and tubby (tub): two autosomal
recessive mutations causing obesity syndromes in the mouse. J. Hered. 81,
424-427 (1990).
786. Vavra, J.J., Deboer, C., Dietz, A., Hanka, L.J. & Sokolski, W.T.
Streptozotocin, a new antibacterial antibiotic. Antibiot Annu 7, 230-5
(1959).
787. Szkudelski, T. The mechanism of alloxan and streptozotocin action in B
cells of the rat pancreas. Physiol Res 50, 537-546 (2001).
788. Lenzen, S. The mechanisms of alloxan- and streptozotocin-induced
diabetes. Diabetologia 51, 216-226 (2008).
789. McEvoy, R.C., Andersson, J., Sandler, S. & Hellerström, C. Multiple low-
dose streptozotocin-induced diabetes in the mouse. Evidence for
216

stimulation of a cytotoxic cellular immune response against an insulin-
producing beta cell line. J. Clin. Invest 74, 715-722 (1984).
790. Oberg, K. Chemotherapy and biotherapy in the treatment of
neuroendocrine tumours. Ann Oncol 12, S111–S114 (2001).
791. Rakieten, N., Rakieten, M.L. & Nadkarni, M.V. Studies on the diabetogenic
action of streptozotocin (NSC-37917). Cancer Chemother Rep 29, 91-98
(1963).
792. Dunn, J.S., Sheehan, H.L. & McLetchie, N.G.B. Necrosis of islets of
Langerhans produced experimentally. The Lancet 241, 484-487 (1943).
793. Isermann, B. et al. Activated protein C protects against diabetic
nephropathy by inhibiting endothelial and podocyte apoptosis. Nat. Med
13, 1349-1358 (2007).
794. Burcelin, R., Kande, J., Ricquier, D. & Girard, J. Changes in uncoupling
protein and GLUT4 glucose transporter expressions in interscapular brown
adipose tissue of diabetic rates: relative roles of hyperglycaemia and
hypoinsulinaemia. Biochem J 291, 109–113 (1993).
795. Tancrède, G., Rousseau-Migneron, S. & Nadeau, A. Long-term changes in
the diabetic state induced by different doses of streptozotocin in rats. Br J
Exp Pathol 64, 117-23 (1983).
796. Hummel, K.P., Dickie, M.M. & Coleman, D.L. Diabetes, a new mutation in
the mouse. Science 153, 1127-1128 (1966).
797. Toye, A.A. et al. A genetic and physiological study of impaired glucose
homeostasis control in C57BL/6J mice. Diabetologia 48, 675-686 (2005).
798. Freeman, H.C., Hugill, A., Dear, N.T., Ashcroft, F.M. & Cox, R.D. Deletion
of nicotinamide nucleotide transhydrogenase: a new quantitive trait locus
accounting for glucose intolerance in C57BL/6J mice. Diabetes 55, 2153-
2156 (2006).
799. Huang, T.-T. et al. Genetic modifiers of the phenotype of mice deficient in
mitochondrial superoxide dismutase. Human Molecular Genetics 15, 1187
-1194 (2006).
800. Golde, W.T., Gollobin, P. & Rodriguez, L.L. A rapid, simple, and humane
method for submandibular bleeding of mice using a lancet. Lab Anim (NY)
34, 40–43 (2005).
217

801. Deakin, A.M., Payne, A.N. & Blackwell, G.J. Role of potassium channels in
the regulation of cytokine release from THP-1 cells. Inflammation
Research 41, C188-C190 (1994).
802. Laliberte, R.E., Eggler, J. & Gabel, C.A. ATP treatment of human
monocytes promotes caspase-1 maturation and externalization. J. Biol.
Chem. 274, 36944-36951 (1999).
803. van der Poll, T. & Opal, S.M. Host-pathogen interactions in sepsis. Lancet
Infect Dis 8, 32-43 (2008).
804. Hamon, Y. et al. Interleukin-1beta secretion is impaired by inhibitors of the
Atp binding cassette transporter, ABC1. Blood 90, 2911-2915 (1997).
805. Ponte-Sucre, A., Campos, Y., Vázquez, J., Moll, H. & Mendoza-León, A.
Sensitivity of Leishmania spp. to glibenclamide and 4-aminopyridine: a
tool for the study of drug resistance development. Mem. Inst. Oswaldo
Cruz 92, 601-606 (1997).
806. Seltzer, H.S. A summary of criticisms of the findings and conclusions of the
University Group Diabetes Program (UGDP). Diabetes 21, 976-979 (1972).
807. Kilo, C., Miller, J.P. & Williamson, J.R. The crux of the UGDP. Spurious
results and biologically inappropriate data analysis. Diabetologia 18, 179-
185 (1980).
808. Meinert, C.L. Clinical trials: Design, Conduct, and Analysis. (Oxford
University Press: Oxford, United Kingdom, 1986).
809. Feinglos, M.N. & Bethel, M.A. Therapy of type 2 diabetes, cardiovascular
death, and the UGDP. Am. Heart J 138, S346-352 (1999).
810. Khalangot, M., Tronko, M., Kravchenko, V. & Kovtun, V. Glibenclamide-
related excess in total and cardiovascular mortality risks: data from large
Ukrainian observational cohort study. Diabetes Res. Clin. Pract. 86, 247-
253 (2009).
811. Monami, M. et al. Are sulphonylureas all the same? A cohort study on
cardiovascular and cancer-related mortality. Diabetes Metab. Res. Rev. 23,
479-484 (2007).
812. Schramm, T.K. et al. Mortality and cardiovascular risk associated with
different insulin secretagogues compared with metformin in type 2
diabetes, with or without a previous myocardial infarction: a nationwide
study. Eur. Heart J. 32, 1900-1908 (2011).
218

813. Schwartz, T.B. & Curtis L, M. The UGDP controversy: thirty-four years of
contentious ambiguity laid to rest. Perspect Biol Med 47, 564–574 (2004).
814. Tomai, F. et al. Effects of K(ATP) channel blockade by glibenclamide on the
warm-up phenomenon. Eur. Heart J 20, 196-202 (1999).
815. Ovünç, K. Effects of glibenclamide, a K(ATP) channel blocker, on warm-up
phenomenon in type II diabetic patients with chronic stable angina
pectoris. Clin Cardiol 23, 535-539 (2000).
816. Tomai, F. et al. Ischemic preconditioning during coronary angioplasty is
prevented by glibenclamide, a selective ATP-sensitive K+ channel blocker.
Circulation 90, 700-705 (1994).
817. Klepzig, H. et al. Sulfonylureas and ischaemic preconditioning; a double-
blind, placebo-controlled evaluation of glimepiride and glibenclamide. Eur
Heart J 20, 439–46 (1999).
818. Lee, T.-M. & Chou, T.-F. Impairment of myocardial protection in type 2
diabetic patients. J. Clin. Endocrinol. Metab 88, 531-537 (2003).
819. Flynn, D.M. et al. The sulfonylurea glipizide does not inhibit ischemic
preconditioning in anesthetized rabbits. Cardiovasc Drugs Ther 19, 337-
346 (2005).
820. Neiszner, E. et al. Influence of diabetic state and that of different
sulfonylureas on the size of myocardial infarction with and without
ischemic preconditioning in rabbits. Exp Clin Endocrinol Diabetes 110,
212–8 (2002).
821. Hueb, W. et al. Effect of a hypoglycemic agent on ischemic preconditioning
in patients with type 2 diabetes and stable angina pectoris. Coron. Artery
Dis 18, 55-59 (2007).
822. Geisen, K., Végh, A., Krause, E. & Papp, J.G. Cardiovascular effects of
conventional sulfonylureas and glimepiride. Horm. Metab. Res 28, 496-
507 (1996).
823. Rydberg, T., Jonsson, A., Roder, M. & Melander, A. Hypoglycemic activity
of glyburide (glibenclamide) metabolites in humans. Diabetes Care 17,
1026-1030 (1994).
824. Balant, L., Fabre, J., Loutan, L. & Samimi, H. Does 4-trans-hydroxy-
glibenclamide show hypoglycemic activity? Arzneimittelforschung 29, 162-
163 (1979).
219

825. Asplund, K., Wiholm, B.E. & Lithner, F. Glibenclamide-associated
hypoglycaemia: a report on 57 cases. Diabetologia 24, 412-417 (1983).
826. Ravindran, S. et al. Identification of glyburide metabolites formed by
hepatic and placental microsomes of humans and baboons. Biochem.
Pharmacol 72, 1730-1737 (2006).
827. Hill, R.A. et al. Hydroxyl-substituted sulfonylureas as potent inhibitors of
specific [3H]glyburide binding to rat brain synaptosomes. Bioorg. Med.
Chem 11, 2099-2113 (2003).
828. Blicklé, J.F. Meglitinide analogues: a review of clinical data focused on
recent trials. Diabetes Metab 32, 113-120 (2006).
220

Creative Commons Licence
Creative Commons License Attribution-Noncommercial-Share Alike
2.0
THE WORK (AS DEFINED BELOW) IS PROVIDED UNDER THE TERMS OF
THIS CREATIVE COMMONS PUBLIC LICENCE (“CCPL” OR “LICENCE”).
THE WORK IS PROTECTED BY COPYRIGHT AND/OR OTHER APPLICABLE
LAW. ANY USE OF THE WORK OTHER THAN AS AUTHORIZED UNDER
THIS LICENCE OR COPYRIGHT LAW IS PROHIBITED. BY EXERCISING
ANY RIGHTS TO THE WORK PROVIDED HERE, YOU ACCEPT AND AGREE
TO BE BOUND BY THE TERMS OF THIS LICENCE. THE LICENSOR GRANTS
YOU THE RIGHTS CONTAINED HERE IN CONSIDERATION OF YOUR
ACCEPTANCE OF SUCH TERMS AND CONDITIONS.
This Creative Commons England and Wales Public Licence enables You (all
capitalised terms defined below) to view, edit, modify, translate and distribute
Works worldwide, under the terms of this licence, provided that You credit the
Original Author.
“The Licensor” (Gavin Christian Kia Wee Koh) and “You” agree as follows:
1. Definitions
1.1. “Attribution” means acknowledging all the parties who have
contributed to and have rights in the Work or Collective Work under
this Licence.
1.1. “Collective Work” means the Work in its entirety in unmodified form
along with a number of other separate and independent works,
assembled into a collective whole.
1.2. “Derivative Work” means any work created by the editing,
modification, adaptation or translation of the Work in any media
(however a work that constitutes a Collective Work will not be
considered a Derivative Work for the purpose of this Licence). For
the avoidance of doubt, where the Work is a musical composition or
221

sound recording, the synchronization of the Work in timed-relation
with a moving image (“synching”) will be considered a Derivative
Work for the purpose of this Licence.
1.3. “Licence” means this Creative Commons England and Wales Public
Licence agreement.
1.4. “Licence Elements” means the following high-level licence attributes
indicated in the title of this Licence: Attribution, Non-Commercial,
Share-Alike.
1.5. “Non-Commercial” means “not primarily intended for or directed
towards commercial advantage or private monetary compensation”.
The exchange of the Work for other copyrighted works by means of
digital file-sharing or otherwise shall not be considered to be
intended for or directed towards commercial advantage or private
monetary compensation, provided there is no payment of any
monetary compensation in connection with the exchange of
copyrighted works.
1.6. “Original Author” means the individual (or entity) who created the
Work.
1.7. “Work” means the work protected by copyright which is offered
under the terms of this Licence.
1.8. For the purpose of this Licence, when not inconsistent with the
context, words in the singular number include the plural number.
2. Licence Terms
2.1. The Licensor hereby grants to You a worldwide, royalty-free, non-
exclusive, Licence for Non-Commercial use and for the duration of
copyright in the Work.
You may:
• copy the Work;
• create one or more Derivative Works;
• incorporate the Work into one or more Collective Works;
• copy Derivative Works or the Work as incorporated in any
Collective Work; and
222

• publish, distribute, archive, perform or otherwise disseminate
the Work or the Work as incorporated in any Collective Work,
to the public in any material form in any media whether now
known or hereafter created.
HOWEVER, You must not:
• impose any terms on the use to be made of the Work, the
Derivative Work or the Work as incorporated in a Collective
Work that alter or restrict the terms of this Licence or any rights
granted under it or has the effect or intent of restricting the
ability to exercise those rights;
• impose any digital rights management technology on the Work
or the Work as incorporated in a Collective Work that alters or
restricts the terms of this Licence or any rights granted under it
or has the effect or intent of restricting the ability to exercise
those rights;
• sublicense the Work;
• subject the Work to any derogatory treatment as defined in the
FINALLY,
You must:
Copyright, Designs and Patents Act 1988.
• make reference to this Licence (by Uniform Resource Identifier
(URI), spoken word or as appropriate to the media used) on all
copies of the Work and Collective Works published, distributed,
performed or otherwise disseminated or made available to the
public by You;
• recognise the Licensor’s / Original Author’s right of attribution
in any Work and Collective Work that You publish, distribute,
perform or otherwise disseminate to the public and ensure that
You credit the Licensor / Original Author as appropriate to the
media used; and
• to the extent reasonably practicable, keep intact all notices that
refer to this Licence, in particular the URI, if any, that the
Licensor specifies to be associated with the Work, unless such
223

URI does not refer to the copyright notice or licensing
information for the Work.
Additional Provisions for third parties making use of the Work
2.2 Further licence from the Licensor
Each time You publish, distribute, perform or otherwise disseminate
• the Work; or
• any Derivative Work; or
• the Work as incorporated in a Collective Work
the Licensor agrees to offer to the relevant third party making use of
the Work (in any of the alternatives set out above) a licence to use the
Work on the same terms and conditions as granted to You hereunder.
2.3. Further licence from You
Each time You publish, distribute, perform or otherwise disseminate
• a Derivative Work; or
• a Derivative Work as incorporated in a Collective Work
You agree to offer to the relevant third party making use of the Work
(in either of the alternatives set out above) a licence to use the
Derivative Work on any of the following premises:
• a licence on the same terms and conditions as the licence
granted to You hereunder; or
• a later version of the licence granted to You hereunder; or
• any other Creative Commons licence with the same Licence
Elements.
2.4. This Licence does not affect any rights that the User may have under
any applicable law, including fair use, fair dealing or any other legally
recognised limitation or exception to copyright infringement.
2.5. All rights not expressly granted by the Licensor are hereby reserved,
including but not limited to, the exclusive right to collect, whether
individually or via a licensing body, such as a collecting society,
224

oyalties for any use of the Work which results in commercial
advantage or private monetary compensation.
3. Warranties and Disclaimer
Except as required by law, the Work is licensed by the Licensor on an “as
is” and “as available” basis and without any warranty of any kind, either
express or implied.
4. Limit of Liability
Subject to any liability which may not be excluded or limited by law the
Licensor shall not be liable and hereby expressly excludes all liability for
loss or damage howsoever and whenever caused to You.
5. Termination
The rights granted to You under this Licence shall terminate automatically
upon any breach by You of the terms of this Licence. Individuals or entities
who have received Collective Works from You under this Licence, however,
will not have their Licences terminated provided such individuals or
entities remain in full compliance with those Licences.
6. General
6.1. The validity or enforceability of the remaining terms of this
agreement is not affected by the holding of any provision of it to be
invalid or unenforceable.
6.2. This Licence constitutes the entire Licence Agreement between the
parties with respect to the Work licensed here. There are no
understandings, agreements or representations with respect to the
Work not specified here. The Licensor shall not be bound by any
additional provisions that may appear in any communication in any
form.
6.3. A person who is not a party to this Licence shall have no rights under
the Contracts (Rights of Third Parties) Act 1999 to enforce any of its
terms.
225

6.4. This Licence shall be governed by the law of England and Wales and
the parties irrevocably submit to the exclusive jurisdiction of the
Courts of England and Wales.
7. On the role of Creative Commons
7.1. Neither the Licensor nor the User may use the Creative Commons
logo except to indicate that the Work is licensed under a Creative
Commons Licence. Any permitted use has to be in compliance with
the Creative Commons trade mark usage guidelines at the time of use
of the Creative Commons trade mark. These guidelines may be found
on the Creative Commons website or be otherwise available upon
request from time to time.
7.2. Creative Commons Corporation does not profit financially from its
role in providing this Licence and will not investigate the claims of
any Licensor or user of the Licence.
7.3. One of the conditions that Creative Commons Corporation requires
of the Licensor and You is an acknowledgement of its limited role and
agreement by all who use the Licence that the Corporation is not
responsible to anyone for the statements and actions of You or the
Licensor or anyone else attempting to use or using this Licence.
7.4. Creative Commons Corporation is not a party to this Licence, and
makes no warranty whatsoever in connection to the Work or in
connection to the Licence, and in all events is not liable for any loss
or damage resulting from the Licensor’s or Your reliance on this
Licence or on its enforceability.
7.5. USE OF THIS LICENCE MEANS THAT YOU AND THE LICENSOR
EACH ACCEPTS THESE CONDITIONS IN SECTION 7.1, 7.2, 7.3, 7.4
AND EACH ACKNOWLEDGES CREATIVE COMMONS
CORPORATION’S VERY LIMITED ROLE AS A FACILITATOR OF
THE LICENCE FROM THE LICENSOR TO YOU.
226

List of manuscripts
Koh GCKW, Maude RR, Schreiber MF, Limmathurotsakul D, Wiersinga WJ,
Wuthiekanun V, Lee SJ, Mahavanakul W, Chaowagul W, Chierakul W, White NJ,
van der Poll T, Day NPJ, Dougan G, Peacock SJ. Glyburide is anti-inflammatory
and associated with reduced mortality in melioidosis. Clin Infect Dis
2011;52(6):717–725.
Koh GCKW, Peacock SJ, van der Poll T, Wiersinga WJ. The impact of
diabetes on the pathogenesis of sepsis. Eur J Clin Microbiol Infect Dis 2011.
PubMed ID: 21805196. [Epub ahead of print]
Koh GCKW, Bretibach K, Kager LM, de Jong HK, Hoogendijk AJ, Day NPJ,
Peacock SJP, van der Poll T, Steinmetz I, Wiersinga WJ. Glibenclamide inhibits
interleukin-1β secretion in a murine model of melioidosis. [Manuscript in
preparation]
227

Acknowledgements
I have been privileged to work in three countries and at three international
centres of excellence. This work could not have been achieved without help
from, collaboration with, and the support of a vast number of people. My job
was only to provide the thread that stitched together the broad canvases of this
painting.
I wish to thank my supervisors, Professor Sharon J Peacock, Professor
Tom van der Poll, Professor Gordon Dougan, Dr Willem Joost Wiersinga and
Professor Nicholas J P Day for their support and guidance, access to their time
and most especially for allowing me to try their patience. I am extremely
grateful to the Wellcome Trust for funding me, for funding this research, for
funding the units within which I work and without whose support this work
would not have been possible.
The cohort study was initially conceived by Professor Peacock and
Dr Direk Limmathurotsakul (mor Dee), using data previously collected at
Sappasithiprasong Hospital in Ubon Ratchathani by Dr Limmathurotsakul,
Dr Wirongrong Chierakul (p’Kae), Dr Bina Maharajan, Saowanit Getchalarat
(p’Luk), Dr Allen Cheng and Dr Atchriya Hemachandra. Dr Rapeephan
Rattanawongnara Maude (Meen) helped me with retrieving the medical notes
and cleaning the data. Dr Limmathurotsakul and Khun Varinthorn Praikaew
(nong Deaw) were responsible for data entry for this study.
Dr Limmathurotsakul and Dr Sue Lee guided me through the statistical
analysis, and Jonathan Sterne helped me to understand and build the
conceptual hierarchical framework.
The work in Thailand was made possible by the support of Professor
Nick Day, who made available to me the staff and resources to do this study,
including the initial decision to fund me for 12 months and the subsequent
decision to fund me for an additional 6 months when it became clear that I was
not efficient enough to complete the study objectives within the time allotted.
The gene expression study was conceived of by Professor Peacock and
Professor Gordon Dougan. Dr Wirongrong Chierakul (p’Kae),
Dr Limmathurotsakul and Khun Vanaporn Wuthiekanun (p’Lek) critiqued the
228

original study proposal and helped presented it to the Thai ethics committee. I
am grateful to the Wellcome Trust Melioidosis Laboratory staff at
Sappasithiprasong, Khun Gumphol Wongsuvan (p’Gum), Khun Sukalya
Paengmee (nong Ya) and Khun Vanaporn Wuthiekanun (p’Lek) for help
diagnosing and identifying patients for the study, and to Dr Limmathurotsakul,
Dr Rattanawongnara Maude, Khun Jintana Suwanapruek (p’Taew), Dr Ekamol
Tantisattamo (mor Bird), Khun Saowanit Getchalarat (p’Luk),
Khun Pittachayanant Ariyaprasert and Khun Maliwan Hongsuvan (N’ng) for
helping to recruit the patients and to collect the blood specimens. I must
acknowledge the sterling work of the medical and nursing staff of
Sappasithiprasong Hospital in caring for the melioidosis patients under difficult
conditions, particularly Dr Suchart Boonraphun, Dr Prapit Teparrukkul,
Dr Praprut Thanakitcharu, Dr Saran Wanapasanee, Dr Akom Arayawichanont,
Dr Wiriya Chualee and Professor Wipada Chaowagul. I also wish to thank the
staff of Warin Chamrap Hospital, Dr Charoen Sayreerattanakorn,
Khun Runglamei Suddee, Khun Uraiwan Kumsrisuk, Khun Niriatchadaporn
Prowsri, Khun Yupadee Nampanya, Khun Supalak Thanomlab and Dr Suthee
Suddee, who run what is probably the best diabetes clinic in the world.
Dr Rattanawongnara Maude was critically responsible for feeding me while I
was in Thailand, teaching me Thai, chauffeuring me around, negotiating on my
behalf, cleaning up after me and for keeping me sane (thank you, sister).
At the Mahidol-Oxford Tropical Medicine Unit in Bangkok, Kanchana
Pongsawat (p’Phung), Sujitra Sukhapiwatana (p’Jiab), Mondira Sarapark
(p’Ting) and nong Anne were indispensible in sorting out the transport of
people, equipment, money and reagents.
At the Academic Medical Center in Amsterdam, I received guidance on
designing the mouse experiments from Dr Willem Joost Wiersinga and
Professor Tom van der Poll. Mouse experiments are not a one-person job and I
am grateful for the technical expertise of Marieke ten Brink and Joost
Daalhuizen in caring for the animals, inoculating them, teaching me how to
administer drugs and when the time came, to sacrifice them humanely and
efficiently. Sample processing would have been much slower and more boring
without the help of (in alphabetical order) Dana Blok, Daan de Boer, Floor van
den Boogaard, Tijmen Hommes, Arjan Hoogendijk, Katja de Jong, Liesbeth
229

Kager, Anne Jan van der Meer. Special thanks goes to Arjan Hoogendijk, who
introduced me gently into lab work and incisive scientific advice. I also wish to
thank Tassili Weehuizen, Liesbeth Kager, Taijmen Hommes, Arjan Hoogendijk
and Regina de Beer for help cutting wax blocks and staining pathology slides.
For cytokine assays (ELISA and CBA), I have to thank Arjan Hoogendijk,
Danielle Kruijswijk, Miriam van Lieshout, Katja de Jong and Tassili Weehuizen.
For the protein assays, I wish to thank Tim Schuijt and Anne Jan van der Meer.
I wish to thank Arjan Hoogendijk, Alex de Vos, Jan Willem Duitman, Teunis
Geitenbeek and Tim Schuijt for helpful and insightful discussions. Special
thanks go to Ajran Hoogendijk and Karla de Souza Queiroz for food, sanity and
spectacularly good Brazilian coffee. Thanks also to Chai Yew Wei for gourmet
Chinese and Italian cuisine. I am especially grateful to Professor van der Poll for
inviting me to work in Amsterdam with him and for providing the laboratory
space and resource without which this project could not have happened.
At the Wellcome Trust Sanger Institute, I wish to thank Professor Gordon
Dougan for providing the resources, space and facilities that enabled me to do
this project. I wish to thank Fernanda Schreiber for extracting the RNA, and the
staff of the Microarray Facility for the cDNA conversion and microarray
hybridisation (particularly Peter Ellis, Ruben Bautista, Rob Kennedy and
Cordelia Langford). I wish to thank Fernanda Schreiber, Roslin Russell, Keith
James, Mike Smith, Robert Stojnic, Ana Luisa Toribio, Mark Dunning, Wei Shi
for help performing the microarray analysis. Special thanks go to Sophie Palmer
for just making things work.
My brain works better for having people to bounce ideas off; for being a
useful sounding board, therefore, I wish to give special thanks to thank Arjan
Hoogendijk, Hsü Li Yang, Emma Nickerson, Richard Maude and Anton
Ilderton. David Dance provided invaluable help obtaining scanned copies of
otherwise unobtainable papers and conference abstracts, despite being in Laos.
I must also acknowledge the help of Anne Bourgeois-Vignon, Samuel Kerrien
and Christopher Kimber for help translating articles from the original French
and German. And to Anton Ilderton, for listening to me complain, cups of tea,
cheering me up, and unconditional support.
In the course of writing this dissertation, I must acknowledge the help of
Sharon Peacock and Joost Wiersinga for their revisions and suggestions. I also
230

wish to thank my examiners Professors Arthur Kaser and Richard Titball for
their helpful comments. All errors remaining are mine alone.
231